Forums

ribbon cable TDR test

Started by John Larkin October 24, 2019
On a sunny day (Fri, 25 Oct 2019 14:44:14 -0700) it happened John Larkin
<jlarkin@highland_atwork_technology.com> wrote in
<liq6repurv3ga15m4tintrc8hq61g1q1em@4ax.com>:

>It's standard 3M ribbon cable. Wiki says "The ribbon cable was >invented in 1956 by Cicoil Corporation." > >I just couldn't find any data on pulse behavior. > >I should test a much longer chunk, to see some serious drool, then >hack a Spice lossy transmission line model to match. > >I was happy with the rise time over a 2' length.
There are many qualities of ribbon cable, the cheap stuff I got from China has less copper[1], weaker plastic than the flat cables to my old harddisks. Never measured those. [1] it is not copper I think... It all depends, I have cheap ethernet cable from China that does not even have twisted pairs.
On 26/10/2019 02:05, Bill Sloman wrote:
> On Saturday, October 26, 2019 at 2:47:37 AM UTC+11, jla...@highlandsniptechnology.com wrote: >> On Fri, 25 Oct 2019 10:34:50 +0100, Michael Kellett <mk@mkesc.co.uk> >> wrote: >> >>> On 25/10/2019 03:39, Lasse Langwadt Christensen wrote: >>>> fredag den 25. oktober 2019 kl. 00.57.24 UTC+2 skrev John Larkin: > > <snip> > >>> Any one care to comment on this related problem. >>> 8 digital outputs (3.3V CMOS from FPGA (guess)) and other things on 37 >>> way D connector, one ground pin for the digital outputs, at one end of >>> the connector. >>> I have to connect this, via about 30cm of cable, to my box and buffer >>> the signals. >>> The goal is to get the best result (for speed and decent pulses) that I can. >>> >>> Plan A is to try with IDC D plugs on the cable and use ribbon cable with >>> T networks of 0603 parts to be optimised in my box at the other end. >>> >>> Plan B is to use tiny resistors in the cable D connector at the source >>> end and T networks of 0603 parts to be optimised in my box at the other >>> end. Cable to be whatever it takes. >>> >>> Plan C is to have an active board at the source plug. >>> >>> Plan D (doing it right with LVDS drivers in the source box) is too late >>> because the source box has been bought and paid for already. >>> >>> A good result would be decent 20ns pulses at 25MHz rate. >> >> Schmitt triggers are always good. You could leave the transmit end >> alone and lowpass+Schmitt each line on your end. > > This is remarkably bad advice. The problem with fast signals on ribbon cables (and every other kind of transmission line) is reflections. > > If you drive a 1.5nec long transmission line with a device that has a lower output impedance than the transmission line (about 120R in this instance) the amplitude of the signal is doubled when it hits far end (potentially damaging the receiver, or at least putting it into a state where it can behave very strangely). Using a Schmitt trigger receiver won't help. > > If the driver is fast enough to produce a 390psec rise-time, you will see this problem. > > At the driven end, the reflection will come back as the same double height signal after 3nsec, and wont do the driver any good either. > >> First guess, 100 ohms and 50 pF then a Schmitt, 5 ns time constant and >> a partial termination to damp any ringing. > > A 100R resistor as source terminator (at the driving end) should do all that is necessary. > > Futzing around at the receiver end is the sort mindless twiddling that ignorant amateurs go in for. >
Thanks (JL and BS) for all the suggestions. I'll try " Futzing around at the receiver end" first - not because it's good, but because it would be so much nicer to be able to use a ribbon cable with IDC connectors plugged straight into the source connector. One of my (many) concerns re. the source box is that it has just one ground pin for many fast digital outputs. Unless I do lots of fancy work inside the connector most of my signals are adjacent to two other independent signals. I'm hoping that the source box has some kind of series termination between connector and drivers but I don't know yet. If I get the job I'll let you know how it goes. MK -- This email has been checked for viruses by AVG. https://www.avg.com
On 26/10/19 09:17, Michael Kellett wrote:
> On 26/10/2019 02:05, Bill Sloman wrote: >> On Saturday, October 26, 2019 at 2:47:37 AM UTC+11, >> jla...@highlandsniptechnology.com wrote: >>> On Fri, 25 Oct 2019 10:34:50 +0100, Michael Kellett <mk@mkesc.co.uk> >>> wrote: >>> >>>> On 25/10/2019 03:39, Lasse Langwadt Christensen wrote: >>>>> fredag den 25. oktober 2019 kl. 00.57.24 UTC+2 skrev John Larkin: >> >> <snip> >>>> Any one care to comment on this related problem. >>>> 8 digital outputs (3.3V CMOS from FPGA (guess)) and other things on 37 >>>> way D connector, one ground pin for the digital outputs, at one end of >>>> the connector. >>>> I have to connect this, via about 30cm of cable, to my box and buffer >>>> the signals. >>>> The goal is to get the best result (for speed and decent pulses) that I can. >>>> >>>> Plan A is to try with IDC D plugs on the cable and use ribbon cable with >>>> T networks of 0603 parts to be optimised in my box at the other end. >>>> >>>> Plan B is to use tiny resistors in the cable D connector at the source >>>> end and T networks of 0603 parts to be optimised in my box at the other >>>> end. Cable to be whatever it takes. >>>> >>>> Plan C is to have an active board at the source plug. >>>> >>>> Plan D&nbsp; (doing it right with LVDS drivers in the source box) is too late >>>> because the source&nbsp; box has been bought and paid for already. >>>> >>>> A good result would be decent 20ns pulses at 25MHz rate. >>> >>> Schmitt triggers are always good. You could leave the transmit end >>> alone and lowpass+Schmitt each line on your end. >> >> This is remarkably bad advice. The problem with fast signals on ribbon cables >> (and every other kind of transmission line) is reflections. >> >> If you drive a 1.5nec long transmission line with a device that has a lower >> output impedance than the transmission line (about 120R in this instance) the >> amplitude of the signal is doubled when it hits far end (potentially damaging >> the receiver, or at least putting it into a state where it can behave very >> strangely). Using a Schmitt trigger receiver won't help. >> >> If the driver is fast enough to produce a 390psec rise-time, you will see this >> problem. >> >> At the driven end, the reflection will come back as the same double height >> signal after 3nsec, and wont do the driver any good either. >>> First guess, 100 ohms and 50 pF then a Schmitt, 5 ns time constant and >>> a partial termination to damp any ringing. >> >> A 100R resistor as source terminator (at the driving end) should do all that >> is necessary. >> >> Futzing around at the receiver end is the sort mindless twiddling that >> ignorant amateurs go in&nbsp; for. >> > Thanks (JL and BS) for all the suggestions. > > I'll try " Futzing around at the receiver end" first - not because it's good, > but because it would be so much nicer to be able to use a ribbon cable with IDC > connectors plugged straight into the source connector. > > One of my (many) concerns re. the source box is that it has just one ground pin > for many fast digital outputs. Unless I do lots of fancy work inside the > connector most of my signals are adjacent to two other independent signals. > > I'm hoping that the source box has some kind of series termination between > connector and drivers but I don't know yet. > > If I get the job I'll let you know how it goes.
The traditional techniques are source termination or receiver termination, but not both. Source termination inserts a series resistor R at the driver so that the driver's output resistance plus R equals the transmission line impedance, Z. Driver supplies V/2Z current. Receiver termination has several options, each with benefits: 1 single resistor to ground Z=R 2 resistor R1 to Vcc, R2 to gnd, such that R1//R2=Z and the potential divider voltage is the receiver's threshold voltage 3 AC only version 1 inserts a capacitor in series All those require the driver can supply double the current, V/Z. 1 asymmetrically loads the driver. 2 has a constant current through R1+R2. 3 is suboptimum except in special circumstances. A schmitt trigger does no harm in properly terminated transmission lines. Your situation (1 gnd in a 37-way cable) is not going to have a well-defined impedance, and there will be more crosstalk, so a schmitt /might/ make the difference between it working or failing. You might also like to consider the effects of static hitting either the driver or the receiver. That can result in catastrophic damage or more subtle parametric shifts.
On Saturday, October 26, 2019 at 7:45:57 PM UTC+11, Tom Gardner wrote:
> On 26/10/19 09:17, Michael Kellett wrote: > > On 26/10/2019 02:05, Bill Sloman wrote: > >> On Saturday, October 26, 2019 at 2:47:37 AM UTC+11, > >> jla...@highlandsniptechnology.com wrote: > >>> On Fri, 25 Oct 2019 10:34:50 +0100, Michael Kellett <mk@mkesc.co.uk> > >>> wrote: > >>> > >>>> On 25/10/2019 03:39, Lasse Langwadt Christensen wrote: > >>>>> fredag den 25. oktober 2019 kl. 00.57.24 UTC+2 skrev John Larkin: > >> > >> <snip> > >>>> Any one care to comment on this related problem. > >>>> 8 digital outputs (3.3V CMOS from FPGA (guess)) and other things on 37 > >>>> way D connector, one ground pin for the digital outputs, at one end of > >>>> the connector. > >>>> I have to connect this, via about 30cm of cable, to my box and buffer > >>>> the signals. > >>>> The goal is to get the best result (for speed and decent pulses) that I can. > >>>> > >>>> Plan A is to try with IDC D plugs on the cable and use ribbon cable with > >>>> T networks of 0603 parts to be optimised in my box at the other end. > >>>> > >>>> Plan B is to use tiny resistors in the cable D connector at the source > >>>> end and T networks of 0603 parts to be optimised in my box at the other > >>>> end. Cable to be whatever it takes. > >>>> > >>>> Plan C is to have an active board at the source plug. > >>>> > >>>> Plan D&nbsp; (doing it right with LVDS drivers in the source box) is too late > >>>> because the source&nbsp; box has been bought and paid for already. > >>>> > >>>> A good result would be decent 20ns pulses at 25MHz rate. > >>> > >>> Schmitt triggers are always good. You could leave the transmit end > >>> alone and lowpass+Schmitt each line on your end. > >> > >> This is remarkably bad advice. The problem with fast signals on ribbon cables (and every other kind of transmission line) is reflections. > >> > >> If you drive a 1.5nec long transmission line with a device that has a lower > >> output impedance than the transmission line (about 120R in this instance) the amplitude of the signal is doubled when it hits far end (potentially damaging the receiver, or at least putting it into a state where it can behave very strangely). Using a Schmitt trigger receiver won't help. > >> > >> If the driver is fast enough to produce a 390psec rise-time, you will see this problem. > >> > >> At the driven end, the reflection will come back as the same double height > >> signal after 3nsec, and won't do the driver any good either. > >> > >>> First guess, 100 ohms and 50 pF then a Schmitt, 5 ns time constant and > >>> a partial termination to damp any ringing. > >> > >> A 100R resistor as source terminator (at the driving end) should do all that is necessary. > >> > >> Futzing around at the receiver end is the sort mindless twiddling that > >> ignorant amateurs go in&nbsp; for. > >> > > Thanks (JL and BS) for all the suggestions. > > > > I'll try " Futzing around at the receiver end" first - not because it's good, but because it would be so much nicer to be able to use a ribbon cable with IDC connectors plugged straight into the source connector.
It's not nice to run the risk of blowing up the receiver. It's not wise to risk feeding current into the catching diodes on the receiver - the current going into them can have unexpected effects. Even if they works as advertised, not blowing up the receiver is not the same as having it work properly while the inputs go outside the supply rails. A T-network can stop that happening, but it messes up the waveform. Source termination is a lot neater, and only needs one damping resistor per lead.
> > One of my (many) concerns re. the source box is that it has just one ground pin for many fast digital outputs. Unless I do lots of fancy work inside the
connector most of my signals are adjacent to two other independent signals. The paths inside the connectors are pretty short, and there are going to be lots of grounded pins in the immediate vicinity too.
> > I'm hoping that the source box has some kind of series termination between > > connector and drivers but I don't know yet.
The source box should have a source termination resistor between each driver pin and the corresponding pin on the connector. The question is whether it's going to have the right value for your particular ribbon cable - it should be pretty much okay since most ribbon cables are pretty similar but you really need to at the driver and the cable before you can be sure.
> > If I get the job I'll let you know how it goes. > > The traditional techniques are source termination or > receiver termination, but not both.
You can do both. It halves your signal level, but really cuts down reflection at both ends of the cable.
> Source termination inserts a series resistor R at the > driver so that the driver's output resistance plus > R equals the transmission line impedance, Z. Driver > supplies V/2Z current.
But only while the cable is getting charged up (or down).
> Receiver termination has several options, each with benefits: > 1 single resistor to ground Z=R > 2 resistor R1 to Vcc, R2 to gnd, such that R1//R2=Z and > the potential divider voltage is the receiver's > threshold voltage > 3 AC only version 1 inserts a capacitor in series > > All those require the driver can supply double > the current, V/Z.
All the time.
> 1 asymmetrically loads the driver.
ECL is designed to drive precisely such terminating resistors - signal swing is from -0.6V (one) to -1.2V (zero), with the terminating resistor returned to -2V
> 2 has a constant > current through R1+R2.
> 3 is suboptimum except in special circumstances,
But it dissipates less heat.
> A schmitt trigger does no harm in properly terminated > transmission lines. Your situation (1 gnd in a 37-way > cable) is not going to have a well-defined impedance, > and there will be more crosstalk, so a schmitt /might/ > make the difference between it working or failing.
If you link every second cable in the ribbon to that single ground you kill almost all the cross-talk.
> You might also like to consider the effects of static > hitting either the driver or the receiver. That can > result in catastrophic damage or more subtle parametric > shifts.
30 cm of cable is likely to be well enough grounded at both ends to make this an unlikely problem. The electron microscopes at Cambridge Instruments had a 30kV supply for the electron gun, and when that flashed over there were unfortunate side effects. Most of them went away when the gun got a proper 30kV coaxial connector. The flash-over currents stayed inside the coax cable, and didn't show up as brief hundred ampere currents in inconvenient places. -- Bill Sloman, Sydney
On Sat, 26 Oct 2019 09:17:35 +0100, Michael Kellett <mk@mkesc.co.uk>
wrote:

>On 26/10/2019 02:05, Bill Sloman wrote: >> On Saturday, October 26, 2019 at 2:47:37 AM UTC+11, jla...@highlandsniptechnology.com wrote: >>> On Fri, 25 Oct 2019 10:34:50 +0100, Michael Kellett <mk@mkesc.co.uk> >>> wrote: >>> >>>> On 25/10/2019 03:39, Lasse Langwadt Christensen wrote: >>>>> fredag den 25. oktober 2019 kl. 00.57.24 UTC+2 skrev John Larkin: >> >> <snip> >> >>>> Any one care to comment on this related problem. >>>> 8 digital outputs (3.3V CMOS from FPGA (guess)) and other things on 37 >>>> way D connector, one ground pin for the digital outputs, at one end of >>>> the connector. >>>> I have to connect this, via about 30cm of cable, to my box and buffer >>>> the signals. >>>> The goal is to get the best result (for speed and decent pulses) that I can. >>>> >>>> Plan A is to try with IDC D plugs on the cable and use ribbon cable with >>>> T networks of 0603 parts to be optimised in my box at the other end. >>>> >>>> Plan B is to use tiny resistors in the cable D connector at the source >>>> end and T networks of 0603 parts to be optimised in my box at the other >>>> end. Cable to be whatever it takes. >>>> >>>> Plan C is to have an active board at the source plug. >>>> >>>> Plan D (doing it right with LVDS drivers in the source box) is too late >>>> because the source box has been bought and paid for already. >>>> >>>> A good result would be decent 20ns pulses at 25MHz rate. >>> >>> Schmitt triggers are always good. You could leave the transmit end >>> alone and lowpass+Schmitt each line on your end. >> >> This is remarkably bad advice. The problem with fast signals on ribbon cables (and every other kind of transmission line) is reflections. >> >> If you drive a 1.5nec long transmission line with a device that has a lower output impedance than the transmission line (about 120R in this instance) the amplitude of the signal is doubled when it hits far end (potentially damaging the receiver, or at least putting it into a state where it can behave very strangely). Using a Schmitt trigger receiver won't help. >> >> If the driver is fast enough to produce a 390psec rise-time, you will see this problem. >> >> At the driven end, the reflection will come back as the same double height signal after 3nsec, and wont do the driver any good either. >> >>> First guess, 100 ohms and 50 pF then a Schmitt, 5 ns time constant and >>> a partial termination to damp any ringing. >> >> A 100R resistor as source terminator (at the driving end) should do all that is necessary. >> >> Futzing around at the receiver end is the sort mindless twiddling that ignorant amateurs go in for. >> >Thanks (JL and BS) for all the suggestions. > >I'll try " Futzing around at the receiver end" first - not because it's >good, but because it would be so much nicer to be able to use a ribbon >cable with IDC connectors plugged straight into the source connector.
Source termination would be great. It reduces the drive currents, which may be a benefit with your grounding. That would involve adding, say, 100 ohms in series with each line at the source, and maybe adding a cap to ground and a Schmitt at the receive end. But you don't know if the source connector is actually close to the source! A series termination midway in a transmission line can get messy. You should probably experiment with the real source and an oscilloscope and some parts. Or find out what's in the source box. LT Spice transmission line models work pretty well.
> >One of my (many) concerns re. the source box is that it has just one >ground pin for many fast digital outputs. Unless I do lots of fancy work >inside the connector most of my signals are adjacent to two other >independent signals.
One shared ground is scary. Can you use a cable shield as another ground?
> >I'm hoping that the source box has some kind of series termination >between connector and drivers but I don't know yet.
Ask! It's a very reasonable question.
> >If I get the job I'll let you know how it goes. > >MK
If it's a big job, adding a little driver board inside a connector shell wouldn't be a big deal. -- John Larkin Highland Technology, Inc lunatic fringe electronics
On 26/10/19 15:36, Bill Sloman wrote:
> On Saturday, October 26, 2019 at 7:45:57 PM UTC+11, Tom Gardner wrote: >> On 26/10/19 09:17, Michael Kellett wrote: >>> On 26/10/2019 02:05, Bill Sloman wrote: >>>> On Saturday, October 26, 2019 at 2:47:37 AM UTC+11, >>>> jla...@highlandsniptechnology.com wrote: >>>>> On Fri, 25 Oct 2019 10:34:50 +0100, Michael Kellett <mk@mkesc.co.uk> >>>>> wrote: >>>>> >>>>>> On 25/10/2019 03:39, Lasse Langwadt Christensen wrote: >>>>>>> fredag den 25. oktober 2019 kl. 00.57.24 UTC+2 skrev John >>>>>>> Larkin: >>>> >>>> <snip> >>>>>> Any one care to comment on this related problem. 8 digital outputs >>>>>> (3.3V CMOS from FPGA (guess)) and other things on 37 way D >>>>>> connector, one ground pin for the digital outputs, at one end of >>>>>> the connector. I have to connect this, via about 30cm of cable, to >>>>>> my box and buffer the signals. The goal is to get the best result >>>>>> (for speed and decent pulses) that I can. >>>>>> >>>>>> Plan A is to try with IDC D plugs on the cable and use ribbon cable >>>>>> with T networks of 0603 parts to be optimised in my box at the >>>>>> other end. >>>>>> >>>>>> Plan B is to use tiny resistors in the cable D connector at the >>>>>> source end and T networks of 0603 parts to be optimised in my box >>>>>> at the other end. Cable to be whatever it takes. >>>>>> >>>>>> Plan C is to have an active board at the source plug. >>>>>> >>>>>> Plan D (doing it right with LVDS drivers in the source box) is too >>>>>> late because the source box has been bought and paid for already. >>>>>> >>>>>> A good result would be decent 20ns pulses at 25MHz rate. >>>>> >>>>> Schmitt triggers are always good. You could leave the transmit end >>>>> alone and lowpass+Schmitt each line on your end. >>>> >>>> This is remarkably bad advice. The problem with fast signals on ribbon >>>> cables (and every other kind of transmission line) is reflections. >>>> >>>> If you drive a 1.5nec long transmission line with a device that has a >>>> lower output impedance than the transmission line (about 120R in this >>>> instance) the amplitude of the signal is doubled when it hits far end >>>> (potentially damaging the receiver, or at least putting it into a state >>>> where it can behave very strangely). Using a Schmitt trigger receiver >>>> won't help. >>>> >>>> If the driver is fast enough to produce a 390psec rise-time, you will >>>> see this problem. >>>> >>>> At the driven end, the reflection will come back as the same double >>>> height signal after 3nsec, and won't do the driver any good either. >>>> >>>>> First guess, 100 ohms and 50 pF then a Schmitt, 5 ns time constant >>>>> and a partial termination to damp any ringing. >>>> >>>> A 100R resistor as source terminator (at the driving end) should do all >>>> that is necessary. >>>> >>>> Futzing around at the receiver end is the sort mindless twiddling that >>>> ignorant amateurs go in for. >>>> >>> Thanks (JL and BS) for all the suggestions. >>> >>> I'll try " Futzing around at the receiver end" first - not because it's >>> good, but because it would be so much nicer to be able to use a ribbon >>> cable with IDC connectors plugged straight into the source connector. > > It's not nice to run the risk of blowing up the receiver. > > It's not wise to risk feeding current into the catching diodes on the > receiver - the current going into them can have unexpected effects. Even if > they works as advertised, not blowing up the receiver is not the same as > having it work properly while the inputs go outside the supply rails. > > A T-network can stop that happening, but it messes up the waveform. > > Source termination is a lot neater, and only needs one damping resistor per > lead. > >>> One of my (many) concerns re. the source box is that it has just one >>> ground pin for many fast digital outputs. Unless I do lots of fancy work >>> inside the > connector most of my signals are adjacent to two other independent signals. > > The paths inside the connectors are pretty short, and there are going to be > lots of grounded pins in the immediate vicinity too. > >>> I'm hoping that the source box has some kind of series termination >>> between connector and drivers but I don't know yet. > > The source box should have a source termination resistor between each driver > pin and the corresponding pin on the connector. > > The question is whether it's going to have the right value for your > particular ribbon cable - it should be pretty much okay since most ribbon > cables are pretty similar but you really need to at the driver and the cable > before you can be sure. > >>> If I get the job I'll let you know how it goes. >> >> The traditional techniques are source termination or receiver termination, >> but not both. > > You can do both. It halves your signal level, but really cuts down reflection > at both ends of the cable.
In most cases you "throwing away" voltage swing is suboptimum. In special, well-designed, cases it might be beneficial, but clearly the OP's case doesn't fall into that category.
>> Source termination inserts a series resistor R at the driver so that the >> driver's output resistance plus R equals the transmission line impedance, >> Z. Driver supplies V/2Z current. > > But only while the cable is getting charged up (or down).
Yes, but that's long enough for problems w.r.t. ground bounce and crosstalk.
>> Receiver termination has several options, each with benefits: 1 single >> resistor to ground Z=R 2 resistor R1 to Vcc, R2 to gnd, such that R1//R2=Z >> and the potential divider voltage is the receiver's threshold voltage 3 AC >> only version 1 inserts a capacitor in series >> >> All those require the driver can supply double the current, V/Z. > > All the time.
Of course.
>> 1 asymmetrically loads the driver. > > ECL is designed to drive precisely such terminating resistors - signal swing > is from -0.6V (one) to -1.2V (zero), with the terminating resistor returned > to -2V
ECL and "derivatives" are nice to work with, but I doubt they are relevant in this case.
>> 2 has a constant current through R1+R2. > >> 3 is suboptimum except in special circumstances, > > But it dissipates less heat.
That's about the only benefit.
>> A schmitt trigger does no harm in properly terminated transmission lines. >> Your situation (1 gnd in a 37-way cable) is not going to have a >> well-defined impedance, and there will be more crosstalk, so a schmitt >> /might/ make the difference between it working or failing. > > If you link every second cable in the ribbon to that single ground you kill > almost all the cross-talk.
Yup, but that isn't the case here :(
>> You might also like to consider the effects of static hitting either the >> driver or the receiver. That can result in catastrophic damage or more >> subtle parametric shifts. > > 30 cm of cable is likely to be well enough grounded at both ends to make this > an unlikely problem.
There's a minor industry devoted to just that problem! If the cable is entirely within a cabinet, and basic anti-static precautions are taken when the cable isn't connected, then I wouldn't be too concerned. I don't know whether that is the case here.
> The electron microscopes at Cambridge Instruments had a 30kV supply for the > electron gun, and when that flashed over there were unfortunate side > effects. > > Most of them went away when the gun got a proper 30kV coaxial connector. The > flash-over currents stayed inside the coax cable, and didn't show up as brief > hundred ampere currents in inconvenient places.
Yes.
On Sunday, October 27, 2019 at 6:02:27 AM UTC+11, Tom Gardner wrote:
> On 26/10/19 15:36, Bill Sloman wrote: > > On Saturday, October 26, 2019 at 7:45:57 PM UTC+11, Tom Gardner wrote: > >> On 26/10/19 09:17, Michael Kellett wrote: > >>> On 26/10/2019 02:05, Bill Sloman wrote: > >>>> On Saturday, October 26, 2019 at 2:47:37 AM UTC+11, > >>>> jla...@highlandsniptechnology.com wrote: > >>>>> On Fri, 25 Oct 2019 10:34:50 +0100, Michael Kellett <mk@mkesc.co.uk> > >>>>> wrote: > >>>>> > >>>>>> On 25/10/2019 03:39, Lasse Langwadt Christensen wrote: > >>>>>>> fredag den 25. oktober 2019 kl. 00.57.24 UTC+2 skrev John > >>>>>>> Larkin:
<snip>
> >> The traditional techniques are source termination or receiver termination, > >> but not both. > > > > You can do both. It halves your signal level, but really cuts down reflection at both ends of the cable. > > In most cases you "throwing away" voltage swing is suboptimum.
Obviously, but there are situations where minimising reflections is essential, and trumps every other consideration.
> In special, well-designed, cases it might be beneficial, but > clearly the OP's case doesn't fall into that category.
Whoever said it did? You were talking about "traditional techniques", not the OP's situation, and termination at both ends of cable is a traditional technique in certain particularly demanding situations. You may not have access to a complete set of traditional techniques ...
> >> Source termination inserts a series resistor R at the driver so that the > >> driver's output resistance plus R equals the transmission line impedance, > >> Z. Driver supplies V/2Z current. > > > > But only while the cable is getting charged up (or down). > > Yes, but that's long enough for problems w.r.t. ground bounce > and crosstalk.
But can reduce power supply load and heat dissipation. Coping with V/Z all the time isn't usually necessary.
> >> Receiver termination has several options, each with benefits: 1 single > >> resistor to ground Z=R 2 resistor R1 to Vcc, R2 to gnd, such that R1//R2=Z > >> and the potential divider voltage is the receiver's threshold voltage 3 AC > >> only version 1 inserts a capacitor in series > >> > >> All those require the driver can supply double the current, V/Z. > > > > All the time. > > Of course. > > >> 1 asymmetrically loads the driver. > > > > ECL is designed to drive precisely such terminating resistors - signal swing > > is from -0.6V (one) to -1.2V (zero), with the terminating resistor returned > > to -2V > > ECL and "derivatives" are nice to work with, but I doubt > they are relevant in this case.
Probably not, but they do show up frequently in the kinds of high-speed circuits that need this sort of attention.
> >> 2 has a constant current through R1+R2. > > > >> 3 is suboptimum except in special circumstances, > > > > But it dissipates less heat. > > That's about the only benefit.
But it can be a useful - sometimes vital - benefit.
> >> A schmitt trigger does no harm in properly terminated transmission lines. > >> Your situation (1 gnd in a 37-way cable) is not going to have a > >> well-defined impedance, and there will be more crosstalk, so a schmitt > >> /might/ make the difference between it working or failing. > > > > If you link every second cable in the ribbon to that single ground you kill > > almost all the cross-talk. > > Yup, but that isn't the case here :(
Michael Kellett just said that he had a 37-way ribbon cable. That should be enough ways to let him devote 15 ways to burying each of the eight data lines between grounded wires. He'd be mad not to.
> >> You might also like to consider the effects of static hitting either the > >> driver or the receiver. That can result in catastrophic damage or more > >> subtle parametric shifts. > > > > 30 cm of cable is likely to be well enough grounded at both ends to make this an unlikely problem. > > There's a minor industry devoted to just that problem!
Cleaning up after people who didn't think about it in advance.
> If the cable is entirely within a cabinet, and basic anti-static > precautions are taken when the cable isn't connected, then > I wouldn't be too concerned. I don't know whether that is the > case here.
It ought to be. Somebody who thinks hard enough to post questions here can probably be relied on to have done their homework.
> > The electron microscopes at Cambridge Instruments had a 30kV supply for the > > electron gun, and when that flashed over there were unfortunate side > > effects. > > > > Most of them went away when the gun got a proper 30kV coaxial connector. The flash-over currents stayed inside the coax cable, and didn't show up as brief hundred ampere currents in inconvenient places. > > Yes.
The odd part about the story was that the proper 30kV coaxial connector got bought and fitted as a sort of make-the-product-look-neater exercise. The improvement in flashover performance came as a surprise. The people surprised were remarkably competent, which is why I mention it here from time to time. As patent lawyers say, everything is obvious to the Supreme Court, but lesser mortals have blind spots. -- Bill Sloman, Sydney
On 27/10/19 02:08, Bill Sloman wrote:
> On Sunday, October 27, 2019 at 6:02:27 AM UTC+11, Tom Gardner wrote: >> On 26/10/19 15:36, Bill Sloman wrote: >>> On Saturday, October 26, 2019 at 7:45:57 PM UTC+11, Tom Gardner wrote: >>>> On 26/10/19 09:17, Michael Kellett wrote: >>>>> On 26/10/2019 02:05, Bill Sloman wrote: >>>>>> On Saturday, October 26, 2019 at 2:47:37 AM UTC+11, >>>>>> jla...@highlandsniptechnology.com wrote: >>>>>>> On Fri, 25 Oct 2019 10:34:50 +0100, Michael Kellett >>>>>>> <mk@mkesc.co.uk> wrote: >>>>>>> >>>>>>>> On 25/10/2019 03:39, Lasse Langwadt Christensen wrote: >>>>>>>>> fredag den 25. oktober 2019 kl. 00.57.24 UTC+2 skrev John >>>>>>>>> Larkin: > > <snip> > > >>>> The traditional techniques are source termination or receiver >>>> termination, but not both. >>> >>> You can do both. It halves your signal level, but really cuts down >>> reflection at both ends of the cable. >> >> In most cases you "throwing away" voltage swing is suboptimum. > > Obviously, but there are situations where minimising reflections is > essential, and trumps every other consideration. > >> In special, well-designed, cases it might be beneficial, but clearly the >> OP's case doesn't fall into that category. > > Whoever said it did? You were talking about "traditional techniques", not the > OP's situation, and termination at both ends of cable is a traditional > technique in certain particularly demanding situations. You may not have > access to a complete set of traditional techniques ... > >>>> Source termination inserts a series resistor R at the driver so that >>>> the driver's output resistance plus R equals the transmission line >>>> impedance, Z. Driver supplies V/2Z current. >>> >>> But only while the cable is getting charged up (or down). >> >> Yes, but that's long enough for problems w.r.t. ground bounce and >> crosstalk. > > But can reduce power supply load and heat dissipation. Coping with V/Z all > the time isn't usually necessary. > >>>> Receiver termination has several options, each with benefits: 1 single >>>> resistor to ground Z=R 2 resistor R1 to Vcc, R2 to gnd, such that >>>> R1//R2=Z and the potential divider voltage is the receiver's threshold >>>> voltage 3 AC only version 1 inserts a capacitor in series >>>> >>>> All those require the driver can supply double the current, V/Z. >>> >>> All the time. >> >> Of course. >> >>>> 1 asymmetrically loads the driver. >>> >>> ECL is designed to drive precisely such terminating resistors - signal >>> swing is from -0.6V (one) to -1.2V (zero), with the terminating resistor >>> returned to -2V >> >> ECL and "derivatives" are nice to work with, but I doubt they are relevant >> in this case. > > Probably not, but they do show up frequently in the kinds of high-speed > circuits that need this sort of attention. > >>>> 2 has a constant current through R1+R2. >>> >>>> 3 is suboptimum except in special circumstances, >>> >>> But it dissipates less heat. >> >> That's about the only benefit. > > But it can be a useful - sometimes vital - benefit. > >>>> A schmitt trigger does no harm in properly terminated transmission >>>> lines. Your situation (1 gnd in a 37-way cable) is not going to have a >>>> well-defined impedance, and there will be more crosstalk, so a schmitt >>>> /might/ make the difference between it working or failing. >>> >>> If you link every second cable in the ribbon to that single ground you >>> kill almost all the cross-talk. >> >> Yup, but that isn't the case here :( > > Michael Kellett just said that he had a 37-way ribbon cable. That should be > enough ways to let him devote 15 ways to burying each of the eight data lines > between grounded wires. > > He'd be mad not to.
Well yes. I presume there aren't sufficient spare connections, or that the pinout is pre-ordained. If the latter, I'd try to find a way of mutating the pinout.
>>>> You might also like to consider the effects of static hitting either >>>> the driver or the receiver. That can result in catastrophic damage or >>>> more subtle parametric shifts. >>> >>> 30 cm of cable is likely to be well enough grounded at both ends to make >>> this an unlikely problem. >> >> There's a minor industry devoted to just that problem! > > Cleaning up after people who didn't think about it in advance.
Or where cables will be external and inserted/removed by the ignorant in uncontrolled conditions. USB is the obvious example, but there are many others.
>> If the cable is entirely within a cabinet, and basic anti-static >> precautions are taken when the cable isn't connected, then I wouldn't be >> too concerned. I don't know whether that is the case here. > > It ought to be. Somebody who thinks hard enough to post questions here can > probably be relied on to have done their homework.
Indeed, but there may be other "users" later on.
>>> The electron microscopes at Cambridge Instruments had a 30kV supply for >>> the electron gun, and when that flashed over there were unfortunate side >>> effects. >>> >>> Most of them went away when the gun got a proper 30kV coaxial connector. >>> The flash-over currents stayed inside the coax cable, and didn't show up >>> as brief hundred ampere currents in inconvenient places. >> >> Yes. > > The odd part about the story was that the proper 30kV coaxial connector got > bought and fitted as a sort of make-the-product-look-neater exercise. > > The improvement in flashover performance came as a surprise. > > The people surprised were remarkably competent, which is why I mention it > here from time to time. As patent lawyers say, everything is obvious to the > Supreme Court, but lesser mortals have blind spots.
Well, everyone except me. I'm perfect. My daughter believed that for a while; it was a useful "life lesson" for her.
>> Michael Kellett just said that he had a 37-way ribbon cable. That >> should be >> enough ways to let him devote 15 ways to burying each of the eight >> data lines >> between grounded wires. >> >> He'd be mad not to. > > Well yes. I presume there aren't sufficient > spare connections, or that the pinout is > pre-ordained. > > If the latter, I'd try to find a way of > mutating the pinout. >
Thanks for further suggestions. Seems I didn't make the initial situation quite clear, the source box is a standard thing already bought by my customer. The 37 way D connector has all 37 pins committed. There are 3 gnd pins, all at one end of the D connector. The pins are obviously driven by octal bus drivers (spec says max 35mA per pin and max 70mA per group of 8.) So I have to take the signals from where they tell me (opposite end of D conn from the gnd pins !). MK
On Sunday, October 27, 2019 at 9:11:20 PM UTC+11, Michael Kellett wrote:
> >> Michael Kellett just said that he had a 37-way ribbon cable. That > >> should be > >> enough ways to let him devote 15 ways to burying each of the eight > >> data lines > >> between grounded wires. > >> > >> He'd be mad not to. > > > > Well yes. I presume there aren't sufficient > > spare connections, or that the pinout is > > pre-ordained. > > > > If the latter, I'd try to find a way of > > mutating the pinout. > > > > Thanks for further suggestions. > Seems I didn't make the initial situation quite clear, the source box is > a standard thing already bought by my customer. The 37 way D connector > has all 37 pins committed. There are 3 gnd pins, all at one end of the D > connector. > The pins are obviously driven by octal bus drivers (spec says max 35mA > per pin and max 70mA per group of 8.) > So I have to take the signals from where they tell me (opposite end of D > connector from the ground pins !).
Oops. You may have to find out if you can make it work, and if not, reconcile your customer to a little board to swap the connections around, possibly onto a wider cable. It takes a while to get naive customers to appreciate how expensive it can be to spend money before they know exactly what they need. Something that can't be made to do what they need done is entirely worthless, and the time spent working why it can't be made to work isn't cheap either. -- Bill Sloman, Sydney