Programmable power supply as arbitrary function generator?

[calzap]

I've seen suggestions for using a programmable power supply as a arbitrary function generator, but has anyone actually done it with low/moderately priced units?  It appears it can be done with programmable Aglients and the extensive software libraries Agilent has available, but what about Korads or Attens or Mastechs or even BKs or Insteks or TTIs?  Presumably, the frequency range would be limited, but even a few kHz would be useful.

I've tried to find info on the company-supplied software available for a number of these units and usually can't.  In a couple of cases I was able to find info, and the software is limited to using the computer to set parameters or logging changes, but nothing for programming changes in output over time.  The manuals for most of these units do give the commands for setting parameters, so one could write software to change output over time, but it would be a big job.  I've done a brief search to look for existing third party software to do this to no avail.  One spec that seems to be universally absent from these units is the response time to computer-generated commands.  This spec would be one determinant of the upper frequency limit in using them as computer-driven function generators.

One other difficulty in using a DC PPSU as an arb is that most of them can't reverse polarity, so no true AC output w/o a hardware trick or two.

Comments?

Mike
California, USA

[Rerouter]

the biggest issue you will face is that for a single quadrant power supply, it will need a feature known as active down-programming, essentially a load to pull the output back down with a small or no load attached,

[Psi]

If building a psu yourself high current opamps work well for the output driver.
That way you can source + sink and produce AC signals using +/-

[calzap]

Rerouter,

I guess I need some education on PPSUs.  Don't own one, but am thinking about buying one.  The two PSUs I have now, which are only voltage controllers, drop their voltage instantly if I turn the knob counterclockwise, regardless of load.  Are you saying this isn't the case with PPSUs?

Mike

[Rerouter]

it is the case, but when your wanted a few KHz of bandwidth, the bleed resistor or possibly small current source inside that currently pulls it down will be far from keeping up,

[calzap]

Hmm,  would be interesting if someone who has a PPSU would put a small or no load on it and connect the output to a DSO, then send a command to double or halve the voltage, and see the response time.  I have a DSO, but not a PPSU.

Mike

[amspire]

I've seen suggestions for using a programmable power supply as a arbitrary function generator, but has anyone actually done it with low/moderately priced units?

There are power supplies designed for exactly this purpose. One family you may be able to pick up cheaply is the old Fluke 42xxA  series programmable power supplies. Forgotten all the model numbers - 4210A, 4215A, 4260A, 4265A, 4270A, 4275A and a few more - something like that. These come in different models  -1A or 100mA and up to +/-100V. These have full 4 quadrant outputs (can output positive or negative volts and can sink or source current) and can go up to something like 10KHz with good accuracy. DC accuracy is 0.01%.

The reason why they can be a bargain is they have no controls on the front at all so they leave most people scratching their head wondering what to do with the boxes. They are controlled by a 14bit binary input which is multiplied by an analogue voltage input (if you get the analog input option). So you probably have to do a bit of coding with an Arduino, etc to feed the binary sequence to it, but you end up with a precision, fast 4 quadrant power supply. Some models have BCD binary input and others pure binary. Some have an option for another 2 bits of resolution.

Richard.

[Jay_Diddy_B]

Hi,

The HP 6129C, 6130C and 6131C also have a BCD or binary interface and analog input. They are designed for 4 quadrant operation.

The HP6824A is also worth considering. These are really laboratory power amplifiers.

Jay_Diddy_B

[amspire]

Here are some examples on ebay:
http://www.ebay.com.au/itm/Fluke-4210A-BCD-Programmable-Power-Source-0-to-9-999-V-/230185343280?pt=LH_DefaultDomain_0&hash=item35981d9930
http://www.ebay.com.au/itm/FLUKE-4265A-BINARY-PROGRAMMABLE-POWER-SOURCE-d-/360610852005?pt=LH_DefaultDomain_0&hash=item53f614eca5
http://www.ebay.com.au/itm/Fluke-4250A-BCD-Programmable-Power-Source-/270754699058?pt=LH_DefaultDomain_0&hash=item3f0a3cd732

[calzap]

Thanks for the replies so far.  My main need is for a compact bench DC PSU with digital readout and controls on the front.  I was hoping for some additional functionality via programmability, but it's not essential.   If programming means just setting the output parameters and keeping a log, it's of no use to me at all.  The old Fluke and HP rack-mount boat anchors are just too big despite their original quality and functionality, and I don't want to spend a lot of time programming.

Mike

[amspire]

Thanks for the replies so far.  My main need is for a compact bench DC PSU with digital readout and controls on the front. 

I understand. To summarize, standard power supplies are one quadrant devices. They can only output a single polarity of voltage and current.

To really act as a dynamic signal source, you want two or four quadrant devices. These supplies have the ability to sink as much power as they can source.  A standard power supply has no ability to sink power at all.

Given that a standard power supply is only a one quadrant device, they are not great dynamic output devices. They are really great at the job they do best - supplying DC power.

[AndyC_772]

One spec that seems to be universally absent from these units is the response time to computer-generated commands.  This spec would be one determinant of the upper frequency limit in using them as computer-driven function generators.

I suspect you'd find the upper frequency to be limited by the output smoothing capacitors. Unless a supply is specifically designed to change its output voltage rapidly, I suspect you'd be lucky to get more than a few Hz out of it.

The Agilent 663xB series is designed for this, and even has a special "fast mode" which removes (or at least, reduces) the output capacitance specifically to allow the output voltage to be changed more quickly. The spec sheet says:

Command processing time:
Average
time required for the output voltage
to begin to change following receipt
of digital data is 4 ms for the power
supplies connected directly to the
GPIB. (Display disabled).
Output programming response time:
The rise and fall time (10/90% and
90/10%) of the output voltage is less
than 2 ms (400 ?s in fast mode). The
output voltage change settles within
1 LSB (0.025% x rated voltage) of final
value in less than 6 ms (2 ms in the
fast mode)

Add the time to send commands over the serial interface (I presume you don't have GPIB) and you'll be lucky to get more than a few Hz this way, and any waveform you get will still be a series of steps. And that's with a PSU that's actually designed for this.

[anotherlin]

The Hameg 8142 and its newer revision the 8143 allows arbitrary waveform.
One of the output can even be modulated using an external source.

[calzap]

Oh, I know some of the high-end (>$1000) Agilents, Teks and Hamegs can act as arbs.  Not only is their hardware capable, but software is available and the relevant specs are provided.

My original question was what can low/moderately priced PPSUs do in this regard.  The specs on their time responses, both response time to computer commands and response time of the output to changes, are generally missing.   Many of them use USB 2, so I don't think computer and communication protocol speeds would be limiting.  The capabilities of the company-supplied software with regard to programming a timed sequence of changes are mostly unknown.  The few that are known can't do it.

Even the ability to have a slow, repetitive series of positive square pulses would be useful.  Of course, any other waveform would have to be constructed of steps, but that's how digitally synthesized function generators work.  There are some tricks that might be possible with a 2-channel PPSU to provide positive and negative waveforms.

Mike

[mzzj]

Even the ability to have a slow, repetitive series of positive square pulses would be useful.  Of course, any other waveform would have to be constructed of steps, but that's how digitally synthesized function generators work.  There are some tricks that might be possible with a 2-channel PPSU to provide positive and negative waveforms.

Mike

I think for a very limited budget a pc soundcard will be more usable.
Good quality soundcard is also handy for measuring variety of stuff like checking low-noise voltage regulators.

[larry42]

Comments?

Long story short: you can't in any *useful* way.

 

[calzap]

I don't think a sound card can do 3A at 30V.  I have a very nice Instek function generator for low voltage, low current, high frequency work.

Mike

[mzzj]

I don't think a sound card can do 3A at 30V.  I have a very nice Instek function generator for low voltage, low current, high frequency work.

Mike

In that case buy, build or steal big "audio" amplifier.

I have one "audio" amp in my hobby corner for buffering the function generator output. Can do easily 30V and 3A
If you want you can dc-couple it and upper frequency limit you can probably modify up to 200-400kHz or so.
 

[calzap]

Interesting that using an audio amp with a function generator should be mentioned.  I used this very combo in the '80's to do some research for designing an electroejaculator for animals.  It worked really well, and we published a paper on it in a scientific journal.

Mike

[Smokey]

I used this very combo in the '80's to do some research for designing an electroejaculator for animals.

Did they let you... um... keep the prototype?... heh

Keep in mind that an audio amp is essentially a voltage-controlled voltage-source.  If will deliver whatever current it can to whatever load is connected at whatever voltage you command.  That's why they specify the minimum resistance of the speakers and say not to go below that.  The other thing to keep in mind is that the bandwidth is usually limited to something around 30khz since people can only hear up to 20khz anyway.

If you want a voltage-controlled current-source, you should look at linear servo amplifiers that take current commands.  They will output a current proportional to the command voltage into whatever load is connected.  Servo amps handle reactive loads better (sometimes), but are usually also limited to about 30khz of current bandwidth depending on the load and often have a minimum inductance.

[robrenz]

You might want to check This thread out. This is an extremely versatile bit of kit that can do up to 20 kHz at 20V and 10 amps.

[calzap]

Smokey,

I don't think anyone would want to have an electroejaculator used on them while they were conscious.  Imagine a plastic dildo covered with metal strips or rings  up your butt and then having 18 V of 1 kHz AC applied to the metal.  Our experimental animals (canines) were anesthetized when they were tested. 

I still have some of the prototypes, and if you would like to try it, it can be arranged.  However, I would like Dave to make a video.

Mike

[baljemmett]

I don't think anyone would want to have an electroejaculator used on them while they were conscious.  Imagine a plastic dildo covered with metal strips or rings  up your butt and then having 18 V of 1 kHz AC applied to the metal.

I suspect there's a website out there somewhere that caters to exactly that 'niche interest'

[C]

In my earlier days, I worked on a crystal controlled reference that speed controlled a an 115 volt motor. a very big off the shelf audio was used.