Current favorite bench DC power supply?

[IDEngineer]

I've been putting this off for too long. I'm about to buy a Siglent SDG2042X and figure I'll get a bench supply at the same time.

My most important requirement is that it goes down to a TRUE 0.0VDC. I've got a variable supply but it only goes down to ~0.7VDC, which is still too much for some analog work.

Clean DC out is a must. Multiple independent outputs would be nice but not a requirement; ditto a fancy LCD display. Need at least 15V max out, 1-2 amps is plenty.

There's everything from <$100 Chinese unknowns on eBay to multi-kilobuck supplies from Keysight, etc. I'm thinking a few hundred dollars feels just about right.

Any experience out there with today's offerings? Thanks!

[bd139]

If I could get one I'd grab a second hand HP/Agilent/Keysight E3611A. The things take up little space, are very reliable, no nasty load transients etc and there are tonnes of them available in the US for not much cash. You either get a Chinese POS with a fancy display or a basic high quality supply with LEDs. I'd go for the latter. It's more important with power supplies than anything else.

Hardly ever turn up in the UK though so I opted for a Thurlby TS3022S (dual 30v 1a). The display sucks though. LEDs are better.

[Dubbie]

Those Chinese POS you denigrate are actually mostly very nice.
I'm talking about name brands like Rigol and Siglent etc, not the no name aliexpress SMPS.

I'd way rather have a few extra nice features like an informative screen, LXI control and linking outputs rather than some mythical "American" quality voodoo.

[nctnico]

I'm still very tempted to get a Keysight E36312A. I've looked at Rigol's and Siglent's offerings but they just don't get anywhere near it when it comes to accuracy (and there are the usual issues) and I really like the low current measurement feature on the E36312A.

[tautech]

I've been putting this off for too long. I'm about to buy a Siglent SDG2042X and figure I'll get a bench supply at the same time.

My most important requirement is that it goes down to a TRUE 0.0VDC. I've got a variable supply but it only goes down to ~0.7VDC, which is still too much for some analog work.

Clean DC out is a must. Multiple independent outputs would be nice but not a requirement; ditto a fancy LCD display. Need at least 15V max out, 1-2 amps is plenty.

There's everything from <$100 Chinese unknowns on eBay to multi-kilobuck supplies from Keysight, etc. I'm thinking a few hundred dollars feels just about right.

Any experience out there with today's offerings? Thanks!

Those that have the latest Siglent PSU seem pretty happy with it.
Some reading, links and vids for you to check out:
https://www.eevblog.com/forum/testgear/siglent-spd3303x-e/

[trophosphere]

I bought a Keithley 2280S-32-6 and it has been my favorite power supply as I do embedded systems development and like having the ability to immediately see precise dynamic current consumption in graphical form. It is also very good for sensors and analog circuitry requiring a low noise power source with the added benefit of programming levels/ramp times to test latchup and glitches. There is also the ability to discharge a power source but I haven't used it. It is on though expensive side though.

[DaJMasta]

I'm still very tempted to get a Keysight E36312A. I've looked at Rigol's and Siglent's offerings but they just don't get anywhere near it when it comes to accuracy (and there are the usual issues) and I really like the low current measurement feature on the E36312A.

They're sort of in a different class, even the upgraded SPD3303X or DP832 are half the price of the Keysight, and more than double the output.  There are certainly applications that can require very fine regulation and very low noise.... but how often are you going to be powering them on your bench without any local regulation or power conditioning on board?  If I power up an RF board or a ADC board that's very sensitive to noise... won't the boards have local regulation or LC filters on the rails or both?  Won't they be operating in a noisy environment when uncovered anyways?  A quick bench test shows that a couple of power resistors in an unshielded rats nest shows me <7mV voltage error including the initial warm up of the supply and ~15mV RMS of noise with the 20M bandwidth limit in place on the scope, which is almost all introduced in the wiring (34461A and DS4024 measuring a SPD3303X-E), and that's on the "economy" version of the PSU (that's less than 1/4 the price of the Keysight) that only shows you 10s of mV and mA.

On the inexpensive side of that estimate, I think used gear is the right approach, and especially if you're doing analog work, having something with two or more output channels to give you a split rail is probably best... though getting decent capacity on three rails at around $100 is a tough order.  If you're fine with half an amp on the higher voltages, there's the E3630A.  If a single rail is fine, maybe the 66311/66312 fits the bill (but tends to be a bit more).  If you don't mind analog readouts, there are a number of choices in that ballpark.  As the pricepoint increases, you get into Rigol and Siglent's current offerings, the DP711 or 712 as a single rail option with a higher voltage capability, then the SPD3303X-E is another notch pricier but with triple outputs, then the DP832 is another notch and has a fully adjustable third rail.  Both the Siglent and Rigol triple supplies have high resolution options (software only) to give you another digit of set resolution and readout.

From looking around, I haven't been wowed by any of BK's recent offerings, but I imagine they're at least on part with the cheaper supplies and at least have a brand name with potential support behind them.

[IDEngineer]

getting decent capacity on three rails at around $100 is a tough order.

I'm willing to spend a few hundred bucks:

I'm thinking a few hundred dollars feels just about right.

...and the comments so far are giving me several options to consider. I have no problem with used gear in good condition, so I think a trip to eBay with some HP/Agilent/Keysight part numbers is a must.

I did like one of the screenshots of a Siglent supply, though... it implied that the unit plots voltage/amperage on its LCD. I have current probes and such for really accurate measurements, but it might be nice to have a "quickie" display like that. Just another thing to think about.

[DaJMasta]

Sorry, misread but edited before that post - the info is about the same, just framed in a way that actually reflects the initial post better 


Anyways, the SPD3303X and X-E variants as well as the Rigol supplies offer waveform displays, for the Siglent, you press a button to toggle it and can have each channel toggled separately.  I don't see an easy way to scale the graphs (and they will not show fine detail because of how far they're zoomed), but you still have access to changing the setting in graph mode.  Probably sufficient to see detail in a variable load or with a programmed variable configuration, but not to show very small adjustments made by the supply or load.

[IDEngineer]

Probably sufficient to see detail in a variable load or with a programmed variable configuration, but not to show very small adjustments made by the supply or load.

Right - I'm thinking here of large-scale changes such as turning on selectively powered subsystems on an embedded system, that sort of thing.

Now if they wanted to bring out a BNC with voltage or (especially!) current info that could be dropped into a DSO, that would be incredible....

[IDEngineer]

I bought a Keithley 2280S-32-6 and it has been my favorite power supply... It is on though expensive side though.

Love Keithley's stuff. Just looked at that unit based on your recommendation... ouch, that's a pricey^H^H^H^H^H^H really nice unit! A bit more money than I'm able to spend right now though, especially since I'm also picking up a Siglent SDG2042X at the same time.

[tautech]

I did like one of the screenshots of a Siglent supply, though... it implied that the unit plots voltage/amperage on its LCD. I have current probes and such for really accurate measurements, but it might be nice to have a "quickie" display like that. Just another thing to think about.

Don't be misled by the SPD3303X or X-E imagery shown on the Siglent websites, it shows a previous FW version of the display GUI that doesn't have the real-time actual output shown.
And yes, I've asked them to update the images as the GUI doesn't represent the product as it's shipped today.

Websites have very recently beeen updated: 

[JXL]

...and the comments so far are giving me several options to consider. I have no problem with used gear in good condition, so I think a trip to eBay with some HP/Agilent/Keysight part numbers is a must.

Try HP6237B (0-18V @ 1A and +/-20V @ 0.5A) or if you prefer lower voltage/higher current (6V @ 3A), look at it's sibling HP6236B.  If bench space is not an issue you can try the larger Leader LPS-151 or LPS-152.  These last 2 have current limiting and the schematics are almost identical to HP6236B.  I have both the HP6237B and LPS-151, and they are my most used bench supplies.  You can find these supplies for <<$100 on eBay.

[IDEngineer]

Another thread mentioned the Instek GPE-x323 series. Nice clean display and you can get 3+1 channels for under $400. Has separate rotary voltage adjustments for the channels too, whereas some "share" a single rotary control across all channels (at least that's better than the "press the +/- buttons as fast as you can" designs). Have to search for reviews on the Instek stuff, haven't ever used it before.

EDIT: Just found the GPE-4323 four channel unit for $318 at TEquipment, and that's before the EEV discount. Quite compelling....

[Macbeth]

My HP6632B has a ~ -2mV for 0V DC (yes, minus - it can sink as well as source current). If I set it to 0.001V it gives +3mV. These are the steps in its DAC/ADC.

Riglol DP832 CH1 output OFF is ~ 4.5mV, output ON at 0.000 is ~ 4.4mV and 0.001V ~ 5.4mV, the PSU itself reporting 00.006V and 00.007V respectively.

mV DC measured with my Keithley 2000.

[rdl]

These do the job for me. No fan, quiet, and solid as a rock, but some would consider the 10 mA meter resolution a shortcoming. Jellybean parts, so easily fixable.  The E3611A is basically the same except higher voltage at lower current.

There were some revisions, so the later Agilent versions are probably preferable. I know my HP version of the E3610A has a bit of turn-on voltage overshoot, but the Agilent does not.

[steven4601]

For the past 3 or 4 years I have been enjoying various 1-6 channel output power supplies.

Somehow I default to this nice 2.5 channel digital programmable bench supply for all quick experiments.
2 channels 0-30V 0-3A, and the 3rd is with a switch option for low voltage aux supplies.

What makes it handy is the quick and easy setting of the outputs with 1mv & 1mA resolution.

[plazma]

My favourite atm. is this DIY unit
https://www.eevblog.com/forum/projects/dps5005-portable-mini-lab-psu/msg1285206/#msg1285206

Larger Thurlby Thandar and HP units are in the garage lab.

[nctnico]

I'm still very tempted to get a Keysight E36312A. I've looked at Rigol's and Siglent's offerings but they just don't get anywhere near it when it comes to accuracy (and there are the usual issues) and I really like the low current measurement feature on the E36312A.

They're sort of in a different class, even the upgraded SPD3303X or DP832 are half the price of the Keysight, and more than double the output.  There are certainly applications that can require very fine regulation and very low noise....

It is not just about noise but also about accuracy and resolution. The E36312A's low current measurement can show current accurate to tens of uA.

[IDEngineer]

Spent a lot of time reviewing the E36000A series. Very tempting! But with discount the GPE-4323 is under $300, a full $1100 less than the 5A version of the Keysight. $1100 buys a lot of test equipment, and nothing I'm doing requires the additional precision to be embedded in the power supply. Then again, I tend to keep test equipment for a very long time so it's never a bad investment. (I still have a Phillips PM3214 analog scope on my bench, works great and is still holding calibration!)

Decisions, decisions....

[nctnico]

IMHO the GPE-4323 is also a good choice if you don't need all the bells and whistles but it still has (IIRC) 1mV / 1mA resolution. What I'm missing in my current bench power supply is accurate readout of voltage and current.

[cdev]

Quietness is a very important attribute for a "favorite" bench power supply.

[nctnico]

Quietness is a very important attribute for a "favorite" bench power supply.

I couldn't agree more. Unfortunately many of the interesting older HP/Agilent power supplies have loud fans so they are not very pleasant to work with.

[bd139]

This is why I like the E36xx supplies. Small, quiet, reliable, no fancy shit to go wrong.

[Alex Eisenhut]

They look like direct descendants of the Harrison HP 6284A style supplies.