Reply by John Larkin December 4, 20112011-12-04
On Sun, 04 Dec 2011 10:10:35 -0800, "Mr.CRC"
<crobcBOGUS@REMOVETHISsbcglobal.net> wrote:

>John Larkin wrote: >> We did a 4-channel arb, in VME and as an oem/embedded box. Once we did >> the VME version, sticking it in a box was relatively easy. >> >> http://www.highlandtechnology.com/DSS/T346DS.shtml >> >> It will modulate almost anything onto anything else, and it does have >> programmable bandwidth noise. It will do noisy PWM. It's mostly just >> an FPGA with a lot of math, then some DACs and filters. >> >> If you build a DDS and want a square wave, you have a choice of just >> using the MSB of the phase accumulator (and suffer a full clock of >> jitter) or use the dac to make a sine and run that through a filter >> and comparator (and suffer a DAC LSB of jitter at lower frequencies.) >> What works well is the sine+comparator thing, but keep the DDS >> frequency in its sweet spot and divide down for lower frequencies. >> >> John > >Yes, I have done a little DDS work and grappled with these choices. > >Your box is interesting, but I wish it had a UI. > >Have you thought about building a new one with a UI? > >I'm sure if you have, your next thought must be, "yeah but the Chinese >are making a new one each week, and they are now in the couple $100s range!"
Exactly!
> >This mustn't be a pleasant thought. > >Thank goodness for niche markets!
Double Exactly! The only thing that seems to be rare in arbs is lots of channels. One is usual, two occasionally, seldom more. We do arbs with dozens of synchronized, phase-shifted, gear-ratioed frequencies for jet engine simulation. It's not a huge market. John
Reply by Mr.CRC December 4, 20112011-12-04
John Larkin wrote:
> We did a 4-channel arb, in VME and as an oem/embedded box. Once we did > the VME version, sticking it in a box was relatively easy. > > http://www.highlandtechnology.com/DSS/T346DS.shtml > > It will modulate almost anything onto anything else, and it does have > programmable bandwidth noise. It will do noisy PWM. It's mostly just > an FPGA with a lot of math, then some DACs and filters. > > If you build a DDS and want a square wave, you have a choice of just > using the MSB of the phase accumulator (and suffer a full clock of > jitter) or use the dac to make a sine and run that through a filter > and comparator (and suffer a DAC LSB of jitter at lower frequencies.) > What works well is the sine+comparator thing, but keep the DDS > frequency in its sweet spot and divide down for lower frequencies. > > John
Yes, I have done a little DDS work and grappled with these choices. Your box is interesting, but I wish it had a UI. Have you thought about building a new one with a UI? I'm sure if you have, your next thought must be, "yeah but the Chinese are making a new one each week, and they are now in the couple $100s range!" This mustn't be a pleasant thought. Thank goodness for niche markets! -- _____________________ Mr.CRC crobcBOGUS@REMOVETHISsbcglobal.net SuSE 10.3 Linux 2.6.22.17
Reply by John Larkin December 4, 20112011-12-04
On Sat, 03 Dec 2011 19:18:54 -0800, "Mr.CRC"
<crobcBOGUS@REMOVETHISsbcglobal.net> wrote:

>Hi: > >So few seem to have this. > >I actually budgeted to buy me an Agilent 33522A for 3 grand a year ago, >along with a DSO3000X series scope. I balked at the generator since the >ones I have at work have serious bugs. They have fixed them almost, but >not perfectly. I know they will fix them fully if I send in new >complain-o-grams, but I haven't had time to do that yet. > >Anyway, for $3k of my own money, I don't want it to have bugs. That has >motivated me to just wait and/or scan the lower cost generator market. > >But there are a few features of the Agilent that I really like: > >a. 2 channels >b. PWM modulation -- why do so few have this? >c. Ability to modulate a PWM with noise. This is so cool, because it >let me set up a scheme to simulate shot to shot variations in laser >performance, then develop the data analysis on my LeCroy scope to plot >in real time the relative standard deviation of pulse energy. >d. Ability to externally modulate the PWM. Again, very cool. >e. Settable noise bandwidth! Only the Agilent does this, from what >I've seen. > > >So now I'm having a look at the TTI model TG2511 25MHz Function, >Arbitrary & Pulse Generator: > >http://www.saelig.com/AGB100/AG012.htm > >Which is at least equipped with PWM modulation including external. >Unfortunately, their pulse capabilities don't offer 1ns resolution, >which is why I haven't found a suitable TTI to buy at work where I need >that. > >I don't really want to spend this much though. > >This Rigol DG1022 offers two channels for much less money, but no PWM: > >http://www.saelig.com/AGB100/AG001.htm > >What I really want to do right now is have two synchronized anti-phase >PWM channels, that I can adjust duty cycle to drive a half-bridge. Not >easy to do this for under $3000 it seems. > >Except for Saelig had this Siglent SDG1005 5MHz Waveform Generator on >it's site the other day, and now it disappeared. Seems to offer 2 >channels and PWM. But the jitter spec on the square wave is terrifying: > >http://www.siglent.com/en/product.asp?id=69 > >It's getting hard to keep track of all the new Chinese instrument >makers. I have ignored them until recently, figuring they just weren't >worth the risk. But when you're talking a factor of 10 cheaper, you >might as well buy one just for a spare. > > >Continuing to search for the perfect and cheap signal generator...
We did a 4-channel arb, in VME and as an oem/embedded box. Once we did the VME version, sticking it in a box was relatively easy. http://www.highlandtechnology.com/DSS/T346DS.shtml It will modulate almost anything onto anything else, and it does have programmable bandwidth noise. It will do noisy PWM. It's mostly just an FPGA with a lot of math, then some DACs and filters. If you build a DDS and want a square wave, you have a choice of just using the MSB of the phase accumulator (and suffer a full clock of jitter) or use the dac to make a sine and run that through a filter and comparator (and suffer a DAC LSB of jitter at lower frequencies.) What works well is the sine+comparator thing, but keep the DDS frequency in its sweet spot and divide down for lower frequencies. John
Reply by Mr.CRC December 3, 20112011-12-03
Hi:

So few seem to have this.

I actually budgeted to buy me an Agilent 33522A for 3 grand a year ago,
along with a DSO3000X series scope.  I balked at the generator since the
ones I have at work have serious bugs.  They have fixed them almost, but
not perfectly.  I know they will fix them fully if I send in new
complain-o-grams, but I haven't had time to do that yet.

Anyway, for $3k of my own money, I don't want it to have bugs.  That has
motivated me to just wait and/or scan the lower cost generator market.

But there are a few features of the Agilent that I really like:

a.  2 channels
b.  PWM modulation -- why do so few have this?
c.  Ability to modulate a PWM with noise.  This is so cool, because it
let me set up a scheme to simulate shot to shot variations in laser
performance, then develop the data analysis on my LeCroy scope to plot
in real time the relative standard deviation of pulse energy.
d.  Ability to externally modulate the PWM.  Again, very cool.
e.  Settable noise bandwidth!  Only the Agilent does this, from what
I've seen.


So now I'm having a look at the TTI model TG2511 25MHz Function,
Arbitrary & Pulse Generator:

http://www.saelig.com/AGB100/AG012.htm

Which is at least equipped with PWM modulation including external.
Unfortunately, their pulse capabilities don't offer 1ns resolution,
which is why I haven't found a suitable TTI to buy at work where I need
that.

I don't really want to spend this much though.

This Rigol DG1022 offers two channels for much less money, but no PWM:

http://www.saelig.com/AGB100/AG001.htm

What I really want to do right now is have two synchronized anti-phase
PWM channels, that I can adjust duty cycle to drive a half-bridge.  Not
easy to do this for under $3000 it seems.

Except for Saelig had this Siglent SDG1005 5MHz Waveform Generator on
it's site the other day, and now it disappeared.  Seems to offer 2
channels and PWM.  But the jitter spec on the square wave is terrifying:

http://www.siglent.com/en/product.asp?id=69

It's getting hard to keep track of all the new Chinese instrument
makers.  I have ignored them until recently, figuring they just weren't
worth the risk.  But when you're talking a factor of 10 cheaper, you
might as well buy one just for a spare.


Continuing to search for the perfect and cheap signal generator...


-- 
_____________________
Mr.CRC
crobcBOGUS@REMOVETHISsbcglobal.net
SuSE 10.3 Linux 2.6.22.17