Reply by Joerg November 28, 20162016-11-28
On 2016-11-28 17:04, P E Schoen wrote:
> "Joerg" wrote in message news:ea3cn9Fe1ikU1@mid.individual.net... > >> As suggested before a CMOS version of the 555 will probably go higher >> (check first though) but would have to be ordered. > >> For a quick fix hang a bus driver between 555 and FET gate. Maybe with >> several sections paralleled. That is usually something that can be >> found in the parts bins. > > The real problem appears to be the low voltage from the USB power lines. > The drain voltage on the MOSFET when it is OFF is only about 3.5 volts. > I think the USB switch chosen for this circuit limits current to 300 mA > by becoming a current source when overloaded. > > A CMOS 555 might provide a voltage close to the rail, but the output > current drive is much less than the ~200 mA of the standard part. The > ICM7555 CMOS part has a typical output high of 4.3 volts with 5 volt > supply, at 0.8 mA. It is actually a CMOS inverter with a rated load of 2 > TTL gates. For an 8 mA load, the voltage drops to 2 volts below supply > rail. This is equivalent to a 250 ohm source impedance. The IRL3915 > MOSFET has an input capacitance of almost 2000 pF, and this corresponds > to a TC of about 500 nSec. With a 100 kHz PWM and about 10% duty cycle, > the ON pulse is 1000 nSec, so the TC effectively cuts that in half. This > is also seen in the scope trace. >
That's why I suggested a bus driver. They provide low tens of ohms and you can usually parallel sections inside the same IC. Also, you don't need a FET with 2000pF gate capacitance for this job. -- Regards, Joerg http://www.analogconsultants.com/
Reply by P E Schoen November 28, 20162016-11-28
"Joerg"  wrote in message news:ea3cn9Fe1ikU1@mid.individual.net...

> As suggested before a CMOS version of the 555 will probably go higher > (check first though) but would have to be ordered.
> For a quick fix hang a bus driver between 555 and FET gate. Maybe with > several sections paralleled. That is usually something that can be found > in the parts bins.
The real problem appears to be the low voltage from the USB power lines. The drain voltage on the MOSFET when it is OFF is only about 3.5 volts. I think the USB switch chosen for this circuit limits current to 300 mA by becoming a current source when overloaded. A CMOS 555 might provide a voltage close to the rail, but the output current drive is much less than the ~200 mA of the standard part. The ICM7555 CMOS part has a typical output high of 4.3 volts with 5 volt supply, at 0.8 mA. It is actually a CMOS inverter with a rated load of 2 TTL gates. For an 8 mA load, the voltage drops to 2 volts below supply rail. This is equivalent to a 250 ohm source impedance. The IRL3915 MOSFET has an input capacitance of almost 2000 pF, and this corresponds to a TC of about 500 nSec. With a 100 kHz PWM and about 10% duty cycle, the ON pulse is 1000 nSec, so the TC effectively cuts that in half. This is also seen in the scope trace. Paul
Reply by Joerg November 28, 20162016-11-28
On 2016-11-27 18:40, John Nagle wrote:
> On 11/18/2016 5:19 PM, P E Schoen wrote: >> "John Nagle" wrote in message news:o0nohf$oo9$1@dont-email.me... >> >>> On 11/18/2016 1:11 AM, P E Schoen wrote: >>> >>>> http://www.mouser.com/ds/2/196/irlr3915pbf-936699.pdf >> >>> That's a big help. It's IRLU3915 in through-hole form. >> >>> That simulation is still using the idealized 555 timer, which seems >>> to have unlimited output current. The MOSFET gate current spikes to >>> about 350mA, which won't happen in the real world. So I added a 10 ohm >>> resistor to the sim between the 555 output and the MOSFET input. That >>> limits the current to about 200mA (the real 555 limit) and the >>> switcher still works. Probably don't need that resistor in the real >>> world. >> >> Don't worry too much about the MOSFET gate current spikes. Driving a >> capacitance with a square wave does that. The 555 is not an ideal gate >> driver, but probably adequate. What is important is for the ON voltage >> to be well above the linear threshold. This was not the case previous to >> fixing the 5V supply problems. >> >>>> It is 55V, 30A, with 14 mOhm RdsOn, and will work with as little as >>>> 2V on the gate. > > Well, I substituted that MOSFET, and it doesn't help much. > It will not work with 2V on the gate. It takes 3V in the real > world, and that's consistent with the data sheet. See > Fig. 3 on > > http://www.mouser.com/ds/2/196/irlr3915pbf-936699.pdf > > and note that at 2V, it's starting to turn on, at > 3V it's passing 10X as much current, and at 7V > on the gate, it's fully turned on and passing > 100X as much current as at 2V. This is not a 2V > device. > > As a test, I replaced the 555 timer signal with > a bench waveform generator. This can go to 3V, > but not to 5V, and at 3V, things sort of start to > work. But not very well. There's no inductive kick > from the transformer at turn-off. See this waveform, > which is MOSFET drain to ground, with a 10x probe. > This is with gate drive at 100KHz, 25% duty cycle. > > http://www.aetherltd.com/public/misc/drainvoltage.jpg > > The comparable simulation works much better than the > real world. Much faster rise time at MOSFET turn-off, > and to a higher voltage. > > http://www.aetherltd.com/public/misc/ttydrivermosfettest1.png (screen) > > http://www.aetherltd.com/public/misc/ttydrivermosfettest1.asc (LTspice) > > The blue line is the drain of the MOSFET; the green line is the > gate. The output voltage in the sim ramps up much faster than > in the real world. Both make it to 120VDC eventually, but it's > about 12ms in sim and about 500ms real world (which is too slow > for the application.) I don't understand why it's so much > less effective in the sim. > > The model of the Coilcraft DA2032-AL transformer is probably > inadequate. Here's its data sheet: > > http://www.coilcraft.com/da2032.cfm > > Suggestions? >
As suggested before a CMOS version of the 555 will probably go higher (check first though) but would have to be ordered. For a quick fix hang a bus driver between 555 and FET gate. Maybe with several sections paralleled. That is usually something that can be found in the parts bins. -- Regards, Joerg http://www.analogconsultants.com/
Reply by P E Schoen November 28, 20162016-11-28
"John Nagle"  wrote in message news:o1g5ct$stv$1@dont-email.me...

> On 11/18/2016 5:19 PM, P E Schoen wrote:
>> Don't worry too much about the MOSFET gate current spikes. Driving a >> capacitance with a square wave does that. The 555 is not an ideal gate >> driver, but probably adequate. What is important is for the ON voltage to >> be well above the linear threshold. This was not the case previous to >> fixing the 5V supply problems.
> Well, I substituted that MOSFET, and it doesn't help much. > It will not work with 2V on the gate. It takes 3V in the real world, and > that's consistent with the data sheet. See Fig. 3 on
> http://www.mouser.com/ds/2/196/irlr3915pbf-936699.pdf
Have you fixed the power supply voltage problem? You should be able to get 4.5 volts on the 555 and that should provide at least 2.5-3 volts gate drive. You might be able to tweak the output a bit higher with a pull-up resistor to the 5V supply.
> and note that at 2V, it's starting to turn on, at 3V it's passing 10X as > much current, and at 7V on the gate, it's fully turned on and passing 100X > as much current as at 2V. This is not a 2V device.
It's probably about as good as you're going to get in a TO-220 package. There are many devices in SOIC-8 and SOT-23 packages with 1.8, 1.5, and even lower gate voltages, although the Vce may not be more than 30V. Try IRLML6344, which will conduct 3A with 0.2V drop using a gate voltage of 1.7 volts.
> As a test, I replaced the 555 timer signal with a bench waveform > generator. This can go to 3V, but not to 5V, and at 3V, things sort of > start to work. But not very well. There's no inductive kick from the > transformer at turn-off. See this waveform, which is MOSFET drain to > ground, with a 10x probe. This is with gate drive at 100KHz, 25% duty > cycle.
> http://www.aetherltd.com/public/misc/drainvoltage.jpg
It appears that the supply voltage is only about 3 volts. That is much too low for the circuit to work. It also appears that the signal generator does not have much drive current which is indicated by the slow ON time and the very visible Miller plateau. The delay provides a much reduced effective duty cycle.
> The comparable simulation works much better than the real world. Much > faster rise time at MOSFET turn-off, and to a higher voltage.
> http://www.aetherltd.com/public/misc/ttydrivermosfettest1.png (screen)
> http://www.aetherltd.com/public/misc/ttydrivermosfettest1.asc (LTspice)
> The blue line is the drain of the MOSFET; the green line is the gate. The > output voltage in the sim ramps up much faster than in the real world. > Both make it to 120VDC eventually, but it's about 12ms in sim and about > 500ms real world (which is too slow for the application.) I don't > understand why it's so much less effective in the sim.
> The model of the Coilcraft DA2032-AL transformer is probably inadequate. > Here's its data sheet:
> http://www.coilcraft.com/da2032.cfm
> Suggestions?
The simulation does not adequately model your real-world circuit. The transformer model is probably OK. You should use actual component modules for the diodes. Also, the TPS2030 shown in your original schematic is current-limited to 300 mA. It would help if you could update the schematic, and also perhaps add a proper low-side MOSFET driver. It is also possible to use a BJT such as a MPSA06 or 2N4401. With a 50 ohm base resistor, you will get about 25 mA and with hfe >= 100 you can easily drive 500 mA. Paul
Reply by John Nagle November 27, 20162016-11-27
On 11/18/2016 5:19 PM, P E Schoen wrote:
> "John Nagle" wrote in message news:o0nohf$oo9$1@dont-email.me... > >> On 11/18/2016 1:11 AM, P E Schoen wrote: >> >>> http://www.mouser.com/ds/2/196/irlr3915pbf-936699.pdf > >> That's a big help. It's IRLU3915 in through-hole form. > >> That simulation is still using the idealized 555 timer, which seems >> to have unlimited output current. The MOSFET gate current spikes to >> about 350mA, which won't happen in the real world. So I added a 10 ohm >> resistor to the sim between the 555 output and the MOSFET input. That >> limits the current to about 200mA (the real 555 limit) and the >> switcher still works. Probably don't need that resistor in the real >> world. > > Don't worry too much about the MOSFET gate current spikes. Driving a > capacitance with a square wave does that. The 555 is not an ideal gate > driver, but probably adequate. What is important is for the ON voltage > to be well above the linear threshold. This was not the case previous to > fixing the 5V supply problems. > >>> It is 55V, 30A, with 14 mOhm RdsOn, and will work with as little as >>> 2V on the gate.
Well, I substituted that MOSFET, and it doesn't help much. It will not work with 2V on the gate. It takes 3V in the real world, and that's consistent with the data sheet. See Fig. 3 on http://www.mouser.com/ds/2/196/irlr3915pbf-936699.pdf and note that at 2V, it's starting to turn on, at 3V it's passing 10X as much current, and at 7V on the gate, it's fully turned on and passing 100X as much current as at 2V. This is not a 2V device. As a test, I replaced the 555 timer signal with a bench waveform generator. This can go to 3V, but not to 5V, and at 3V, things sort of start to work. But not very well. There's no inductive kick from the transformer at turn-off. See this waveform, which is MOSFET drain to ground, with a 10x probe. This is with gate drive at 100KHz, 25% duty cycle. http://www.aetherltd.com/public/misc/drainvoltage.jpg The comparable simulation works much better than the real world. Much faster rise time at MOSFET turn-off, and to a higher voltage. http://www.aetherltd.com/public/misc/ttydrivermosfettest1.png (screen) http://www.aetherltd.com/public/misc/ttydrivermosfettest1.asc (LTspice) The blue line is the drain of the MOSFET; the green line is the gate. The output voltage in the sim ramps up much faster than in the real world. Both make it to 120VDC eventually, but it's about 12ms in sim and about 500ms real world (which is too slow for the application.) I don't understand why it's so much less effective in the sim. The model of the Coilcraft DA2032-AL transformer is probably inadequate. Here's its data sheet: http://www.coilcraft.com/da2032.cfm Suggestions? John Nagle
Reply by P E Schoen November 19, 20162016-11-19
I did a bit more on this. I removed the 555 and replaced it with a simple 
pulse generator with 4V output. I replaced the MOSFET with a simple 2N4401 
BJT. I added some parallel capacitance to the transformer, and I tweaked the 
output snubber. Results seem pretty good:

http://enginuitysystems.com/pix/electronics/ttydriver_PES_schematic.png

http://enginuitysystems.com/pix/electronics/ttydriver_PES_waveform.png

http://enginuitysystems.com/pix/electronics/ttydriver22c.asc

I'm sure many more improvements can be made. There is still a lot of high 
frequency ringing on the transformer that perhaps could be reduced. I might 
use a PIC for the PWM, and possibly also for the USB-serial conversion 
rather than the CP2102 module.

Paul 

Reply by P E Schoen November 18, 20162016-11-18
"John Nagle"  wrote in message news:o0nohf$oo9$1@dont-email.me...

> On 11/18/2016 1:11 AM, P E Schoen wrote: > >> http://www.mouser.com/ds/2/196/irlr3915pbf-936699.pdf
> That's a big help. It's IRLU3915 in through-hole form.
> That simulation is still using the idealized 555 timer, which seems to > have unlimited output current. The MOSFET gate current spikes to about > 350mA, which won't happen in the real world. So I added a 10 ohm resistor > to the sim between the 555 output and the MOSFET input. That limits the > current to about 200mA (the real 555 limit) and the switcher still works. > Probably don't need that resistor in the real world.
Don't worry too much about the MOSFET gate current spikes. Driving a capacitance with a square wave does that. The 555 is not an ideal gate driver, but probably adequate. What is important is for the ON voltage to be well above the linear threshold. This was not the case previous to fixing the 5V supply problems.
>> It is 55V, 30A, with 14 mOhm RdsOn, and will work with as little as 2V on >> the gate.
> Sim says 25V, 2.7A max at the MOSFET, so that's within limits.
>> and added a commutating snubber that totally eliminates negative spikes >> and ringing.
> Too much snubbing - the magnet won't release fast enough. Look at the > current through I3. It's above 60mA for the whole OFF period. The energy > in the magnet needs to be dumped in about 2ms or the mechanism won't work.
> The idea is to use just enough snubbing to avoid component damage. The > original 1926-1930 equipment used mechanical contacts which arced when the > circuit opened. Around 1940, RF suppression was added, as a small RC > circuit.
That's my error, and today I saw that the single Zener was acting as a commutating diode and my snubber was actually doing nothing. I was looking at Vselector and not I(L6). I added a silicon diode in series to block the commutating action and now the snubber works as expected. Actually, D6 and the added D8 are probably not needed at all and can be removed. My simulation seems to run awfully slow so I'm not going to run again to verify, but it looks much better now! http://enginuitysystems.com/pix/electronics/ttydriver22d.png I also changed the PWM duty cycle to about 10%. Now the input draws only about 150 mA or 800 mW. I also updated the ASCII file: http://enginuitysystems.com/pix/electronics/ttydriver22c.asc
> I can tell the USB interface to ask for 300mA. No problem.
About half that now ;)
> What's with the note "Draws about 1.7 amps or 6.2 watts?"
That was for one of the previous simulations. Probably when the MOSFET was running in linear mode due to the supply voltage crashing.
> I'm going to try substituting a IRLU3915 for now and take measurements, > then do a new board design. Thanks.
I'd be happy to review your new design before you send off for new boards. This has been an interesting project. I am not particularly "into" antique electronics but I do find them interesting, and I also enjoy working on electro-mechanical contrivances. I have lots of old electronic components and equipment, and I recently toured a local computer museum that has old mechanical computers and various examples of the transition from iron core to solid state memory, relays and vacuum tubes to transistors and ICs, and other technology. They are even restoring a big old linotype machine, and they probably have some teletypes as well. https://museum.syssrc.com/ They may benefit from your work and you are of course welcome to visit. Let me know. I also have a friend who specializes in old electronic stuff, mostly ham radio gear, as well as juke boxes, pinball machines, and scopes and such. He also volunteers at the National Electronics Museum at BWI: https://en.wikipedia.org/wiki/National_Electronics_Museum http://nationalelectronicsmuseum.org/ Paul
Reply by bitrex November 18, 20162016-11-18
On 11/18/2016 01:16 PM, Joerg wrote:
> On 2016-11-18 09:57, Tim Wescott wrote: >> On Wed, 16 Nov 2016 23:33:57 -0800, John Nagle wrote: >> >>> I've been working on a power supply for antique Teletype selector >>> magnets. These need 120VDC 60mA for the first 2ms or so of each bit >>> time, but the sustain voltage needed is only 3-4V. So I've built an >>> isolated boost supply and interface board, all powered from a USB port. >>> The whole project is on Github: >>> >>> https://github.com/John-Nagle/ttyloopdriver >>> >>> where there's an explanation of how it works which doesn't require >>> downloading the whole KiCAD project. >>> >>> Schematic as image: >>> http://www.aetherltd.com/public/misc/loopdriverv1.0.png >>> >>> LTSpice model: http://www.aetherltd.com/public/misc/ttydriver22.asc >>> >>> I've had a board made, and it produced 120VDC >>> briefly, until the IRLB8721 MOSFET blew out. I just used that because >>> it's a common MOSFET. >>> Not surprising - the circuit shows 40V across source and drain, and >>> that's a 30V part. >>> >>> So what to put in there? I've been trying various >>> parts in LTSpice. The AP9465GEM works great in LTSpice, but it's a >>> surface mount part and only 40V, so there's no headroom on voltage. >>> Almost nothing with a rating of 50V or better works in simulation. >>> Turn-off is too slow, or capacitance is too high, and the resulting >>> output voltage peak on the primary is too low. Try loading up the >>> LTSpice model and using different MOSFETs to see this. >>> >>> I don't really know what I'm doing here. I'm >>> aware that this MOSFET application depends strongly on the transient >>> properties of the device. Turning off that transformer primary to get >>> the inductive kick is tricky. I'm trying to struggle through the Art of >>> Electronics chapter on MOSFETs, while hammering on the problem with >>> LTSpice. >>> >>> (I'd like to use a TO-220 form factor, or at least a through-hole part, >>> and get this first board working. After that, I can get a new board >>> made.) >>> >>> (I was asking previously about low duty cycle 555 timer circuits. That >>> part is working fine now. Thanks.) >> >> The selection of TO-220 parts is pretty scant these days. > > > I haven't seen much of an availability problem there. > > >> ... I suspect >> you'll have a much better time of it if you buy DPAK-2 parts and solder >> pigtails onto the lands to make them into sorta-TO-220 parts. Or just >> bite the bullet and start using surface-mount parts throughout. >> > > In this case no problem but I think John wants to have the board > through-hole. TO220 are still widely available and will be for a long > time because bigger supplies than this need heat sinks. You can't > reliably sink more than low single-digit watts into a PCB. I design in > SMT since 1986, was an early adopter but had to use many TO220 parts for > power stages. >
Often in the commercial products from that era I've looked at, like synthesizers, etc. it's all through-hole except for one or two big chips with many pins, like a custom gate array or display controller, which look surprisingly modern by comparison.
Reply by John Nagle November 18, 20162016-11-18
On 11/18/2016 1:11 AM, P E Schoen wrote:
> "John Nagle" wrote in message news:o0ma68$jd6$1@dont-email.me... >
....
> > I could try to analyze what's going on with your most recent efforts, > but I found more problems that are more serious. First off, it's not > going to work with a 5 ohm power source resistance. A USB will supply > close to 5VDC (maybe as low as 4.8) and it will do so until it > approaches the standard 500 mA limit, at which it may throttle back or > shut down completely. So I used your 4.6V with a 0.5 ohm resistance. > That will still mean a drop to about 4.3 volts at 500 mA, but still > workable.
Makes sense. I'd changed that in some of my own simulations.
> Next, I changed the input diode to a Schottky, which drops only about > 0.3 volts rather than almost 1 volt. That was a huge improvement. I also > found a better MOSFET, the IRL3915. It seems that it is actually an > IRLR3915. There are probably even better ones available, but this is > only a dollar and is easily available: > > http://www.mouser.com/ds/2/196/irlr3915pbf-936699.pdf
That's a big help. It's IRLU3915 in through-hole form. That simulation is still using the idealized 555 timer, which seems to have unlimited output current. The MOSFET gate current spikes to about 350mA, which won't happen in the real world. So I added a 10 ohm resistor to the sim between the 555 output and the MOSFET input. That limits the current to about 200mA (the real 555 limit) and the switcher still works. Probably don't need that resistor in the real world.
> It is 55V, 30A, with 14 mOhm RdsOn, and will work with as little as 2V > on the gate.
Sim says 25V, 2.7A max at the MOSFET, so that's within limits.
> and added a commutating snubber that totally eliminates negative > spikes and ringing.
Too much snubbing - the magnet won't release fast enough. Look at the current through I3. It's above 60mA for the whole OFF period. The energy in the magnet needs to be dumped in about 2ms or the mechanism won't work. The idea is to use just enough snubbing to avoid component damage. The original 1926-1930 equipment used mechanical contacts which arced when the circuit opened. Around 1940, RF suppression was added, as a small RC circuit.
> I dropped the frequency to about 40 kHz. Here is the > result: > > http://enginuitysystems.com/pix/electronics/ttydriver22c.png > > Note that this seems to draw only about 300 mA from the supply, or about > 1.6 watts. Fine for a USB source.
I can tell the USB interface to ask for 300mA. No problem. What's with the note "Draws about 1.7 amps or 6.2 watts?" I'm going to try substituting a IRLU3915 for now and take measurements, then do a new board design. Thanks. John Nagle
Reply by Joerg November 18, 20162016-11-18
On 2016-11-18 09:57, Tim Wescott wrote:
> On Wed, 16 Nov 2016 23:33:57 -0800, John Nagle wrote: > >> I've been working on a power supply for antique Teletype selector >> magnets. These need 120VDC 60mA for the first 2ms or so of each bit >> time, but the sustain voltage needed is only 3-4V. So I've built an >> isolated boost supply and interface board, all powered from a USB port. >> The whole project is on Github: >> >> https://github.com/John-Nagle/ttyloopdriver >> >> where there's an explanation of how it works which doesn't require >> downloading the whole KiCAD project. >> >> Schematic as image: >> http://www.aetherltd.com/public/misc/loopdriverv1.0.png >> >> LTSpice model: http://www.aetherltd.com/public/misc/ttydriver22.asc >> >> I've had a board made, and it produced 120VDC >> briefly, until the IRLB8721 MOSFET blew out. I just used that because >> it's a common MOSFET. >> Not surprising - the circuit shows 40V across source and drain, and >> that's a 30V part. >> >> So what to put in there? I've been trying various >> parts in LTSpice. The AP9465GEM works great in LTSpice, but it's a >> surface mount part and only 40V, so there's no headroom on voltage. >> Almost nothing with a rating of 50V or better works in simulation. >> Turn-off is too slow, or capacitance is too high, and the resulting >> output voltage peak on the primary is too low. Try loading up the >> LTSpice model and using different MOSFETs to see this. >> >> I don't really know what I'm doing here. I'm >> aware that this MOSFET application depends strongly on the transient >> properties of the device. Turning off that transformer primary to get >> the inductive kick is tricky. I'm trying to struggle through the Art of >> Electronics chapter on MOSFETs, while hammering on the problem with >> LTSpice. >> >> (I'd like to use a TO-220 form factor, or at least a through-hole part, >> and get this first board working. After that, I can get a new board >> made.) >> >> (I was asking previously about low duty cycle 555 timer circuits. That >> part is working fine now. Thanks.) > > The selection of TO-220 parts is pretty scant these days.
I haven't seen much of an availability problem there.
> ... I suspect > you'll have a much better time of it if you buy DPAK-2 parts and solder > pigtails onto the lands to make them into sorta-TO-220 parts. Or just > bite the bullet and start using surface-mount parts throughout. >
In this case no problem but I think John wants to have the board through-hole. TO220 are still widely available and will be for a long time because bigger supplies than this need heat sinks. You can't reliably sink more than low single-digit watts into a PCB. I design in SMT since 1986, was an early adopter but had to use many TO220 parts for power stages. -- Regards, Joerg http://www.analogconsultants.com/