Bench Power Supply

[Majorbob]

Hi all, sorry if this is the wrong place,

I'm wanting a bench power supply for xmas.  Ideally I'd like dual channel, up to 30V, up to 5A and linear.  I've scoured ebay for second hand ones and it seems to be very thin on the ground.  Budget is quite low, I'd like to keep it under £100 if possible.  And I'm in the UK.  Can anyone shed any light on my search? Most of what's on ebay seems to be low-quality chinese crap.
Thanks.

[Jeroen3]

Most of what's on ebay seems to be low-quality chinese crap.

Yes, sad times. Turn on "used" for the good brands. There are some HP's on it now.
Although, the Korad K3005 ones and rebrands are good enough for their price I would say. Best of the worst, I guess.

[de_light]

I have a Wavecom PS-3005D (Korad rebrand) SMPSU

Pretty happy with this for the price.

Did a quick scope of the unloaded power supply at the output set to 9V.
20M filter on

[radiolistener]

Yes, KORAD KA3005D is the best choice for a universal lab PSU.

But don't expect very clean output, it has a little noise. Probably from dynamic LED indication.

Pulse appears with about 120 Hz (8.3 ms) period and consists of about 34 MHz oscillations

[tggzzz]

There are a couple of Ramco auctions coming up, with many PSUs. Some of those might meet your needs, and hence be worth considering - but don't bid too high!

There's no guarantee with any of the items, so you have to look at the photo carefully and take a chance. Usually, but not always, it is OK.

They ship for £15. Don't forget to add 15.5% and 20% to the hammer price!

[nctnico]

I have a Wavecom PS-3005D (Korad rebrand) SMPSU

Pretty happy with this for the price.

Did a quick scope of the unloaded power supply at the output set to 9V.
20M filter on

Dave did a teardown of the Korad KA3005P PSU which looks similar to the PS-3005D.


At least this isn't a switching PSU.

I Strongly recommend against a switching power supply for doing electronics development because of the noise they can/will inject into your measurements. Don't get me wrong here; switching PSUs are great for delivering lots of power but not for use as a primary bench supply while making (sensitive) measurements.

[Majorbob]

Linear is the only way to go in my view.  The Ramco auction was a great shout, seen some great power supplies on there and hoping to win an Metrix AX503.

[tggzzz]

Linear is the only way to go in my view.  The Ramco auction was a great shout, seen some great power supplies on there and hoping to win an Metrix AX503.

Make sure you are registered well before the auction starts. That process is not instantaneous.

See also the Peaker Pattinson auctions, but beware that at Aunby they store some stuff in a farm barn.

There will be another auction along soon, with more PSUs

[nctnico]

Linear is the only way to go in my view.  The Ramco auction was a great shout, seen some great power supplies on there and hoping to win an Metrix AX503.

That Metrix AX503 doesn't look very compelling to me. Voltage is shown with 1 decimal only. That is pretty much useless for electronics work.

[uski]

Linear is the only way to go in my view.  The Ramco auction was a great shout, seen some great power supplies on there and hoping to win an Metrix AX503.

That Metrix AX503 doesn't look very compelling to me. Voltage is shown with 1 decimal only. That is pretty much useless for electronics work.

I wouldn't say that... many people just need a stable 5V or 12V with some intermediates in between, sometimes.
And with a 100 GBP budget, you can't do miracles.

I would focus on getting a linear PSU with a stable and low-ish noise output.
Then, you can add an external voltmeter. My concern is more about current display, I like to get a precise measurement of the current. But again that can be done externally.

I got myself a Keithley 2280S PSU... 6.5 digit readings straight from the PSU. It's maybe too much . I have an Agilent which reads only 2 digits and I found it is not enough especially for the current.
So a reasonable ground is probably 3 decimals (i.e. down to the mA) and most cheap PSUs don't do that (but there are optons, see my post below). Adding an external meter is probably the way to do it cheap.

[uski]

OP, please consider that your instrument will probably last you for a decade or more. At least that's the experience I have with my instruments.
So it makes sense to "invest" a bit, if you can afford it at all.

With that in mind, have you considered a Siglent PSU such as the SPD3303X-E ?
It is around 300 GBP. You can parallel the 2 outputs for up to 6.4A, or put them in series for up to 64V.
Seems like a very capable power supply, that would probably last you for a while. And having 2 outputs is very useful and it also has a third 2.5/3.3/5V output which is convenient.

There is also the SPD3303X which is sigificantly more expensive, for a higher resolution reading (1mV/1mA instead of 10mA/10mV on the -E)
Some people have hacked the SPD3303X-E to become a SPD3303X. If you are really on a budget, that could be an option...

Rigol also makes the DP832 for instance, I had one and didn't really like the user interface, but it is an option. More expensive however. Can be hacked too.

You know the saying "poor people pay twice". You are probably going to put 100 GBP in a power supply, and unless you are able to find a bargain (which I hope you do !), you might end up buying another more expensive one in a year or two.

Finally... used instruments are fun... but sometimes it is frustrating to spend more time fixing the instrument itself than working on your projects, especially if it is your primary instrument and you don't have another one. Most of the time you can fix it but sometimes you can't and it's frustrating. Getting a new instrument will ensure that you get an instrument you can rely on. I know this from experience, my very first power supply was faultly. I was a student at the time and got a brand new one for christmas... but it was a cheap one... and it was real crap. Now I could afford an expensive power supply and that's what I did. No regrets. So take a look at these Siglents units

[radiolistener]

At least this isn't a switching PSU.

I Strongly recommend against a switching power supply for doing electronics development because of the noise they can/will inject into your measurements. Don't get me wrong here; switching PSUs are great for delivering lots of power but not for use as a primary bench supply while making (sensitive) measurements.

Yes, it has very low noise, much better than very good SMPS. But you can catch some rare small pulses, see my picture above. These pulses are always present on the output even when output is OFF. It disappears when you power off PSU. I didn't find this noise on the spectrum, but you can catch it with oscilloscope. I suspect it comes from the dynamic indication of 7-segment LED.

The LED indicator is very bright, so I'm thinking about reduce LED brightness. Does anyone tried to reduce brightness?

[uski]

At least this isn't a switching PSU.

I Strongly recommend against a switching power supply for doing electronics development because of the noise they can/will inject into your measurements. Don't get me wrong here; switching PSUs are great for delivering lots of power but not for use as a primary bench supply while making (sensitive) measurements.

Yes, it has very low noise, much better than very good SMPS. But you can catch some rare small pulses, see my picture above. These pulses are always present on the output even when output is OFF. It disappears when you power off PSU. I didn't find this noise on the spectrum, but you can catch it with oscilloscope. I suspect it comes from the dynamic indication of 7-segment LED.

Have you done your measurement with any load on the output ? I suspect it could be very high impedance noise. Still not perfect, but it's likely your lab has electrical noise from other things.
In addition if the oscillations are 34MHz, they should be easy to filter out by adding a small LC filter on the line.
But I wonder if it is not oscillation or ringing in your probing setup.

[Stray Electron]

Hi all, sorry if this is the wrong place,

I'm wanting a bench power supply for xmas.  Ideally I'd like dual channel, up to 30V, up to 5A and linear.  I've scoured ebay for second hand ones and it seems to be very thin on the ground.  Budget is quite low, I'd like to keep it under £100 if possible.  And I'm in the UK.  Can anyone shed any light on my search? Most of what's on ebay seems to be low-quality chinese crap.
Thanks.

   Go to Ebay and put in your location and then search for HP or Lambda or other GOOD quality power supply, then tell E-bay to only show USED items, then tell E-bay to show the nearest ones first. Find one that you want and ask the seller if they will allow local pickup (and not charge for it).  Find one that you like, buy it and then go pick it up.  Used power supplies don't sell for much on Ebay, particularly linear supplies because they're so heavy to ship, so buy one locally and save yourself a ton of money.  I recommend the HP supplies because they're GOOD, they're reliable and service info is available for most of them. But choose what you want.

  PS Don't waste your time on some of the ridiculous priced items on Ebay.  I start with the lowest priced items and then work my way up to the higher priced ones until I find something that will do what I want. Buy USED only, don't waste you time on anything non-working unless service manuals for it are available and you WANT a challange. Or buy NEW, but be ready to pay the price!

  If you don't find a suitable one right away, you can save the search and have E-bay E-mail you when one shows up.

   I typically pay between $10 and $20 for older working HP supplies. If you want one with digital displays, expect to pay considerably more but they won't work any better than an old linear supply with analog meters.  I have HP supplies from the 1960s that still work fine. One of the nice thing about the HP supplies is that they used GOOD capacitors and I've never seen a cap fail in one.

[radiolistener]

Have you done your measurement with any load on the output ? I suspect it could be very high impedance noise. Still not perfect, but it's likely your lab has electrical noise from other things.
In addition if the oscillations are 34MHz, they should be easy to filter out by adding a small LC filter on the line.
But I wonder if it is not oscillation or ringing in your probing setup.

It was measured through 1 meter coax cable RG58 with BNC connectors and BNC to banana adapter. There is no load. When I short output terminals with 1 meter cable it doesn't affect this noise. Output ON/OFF also doesn't affect this pulse.

It looks like a little interference from dynamic LED refresh currents. Because when I power off PSU, the pulse disappears.

I tried to connect the same cable to HY1505D PSU and there is no pulses.

On the radio receiver, I don't see any noise on 34-35 MHz.

Just tested KORAD KA3005D and HY1505D output with soundcard input and it looks that KORAD has less noise floor. But it has some noise at about 600 Hz with about -90 dBm (soundcard input is calibrated with 50 Ohm generator).

The first screenshot is KORAD KA3005D output spectrum (the switch moment is pressing ON button).

The second screenshot is HY1505D output spectrum (the switch moment is putting BNC-banana adapter into PSU).

Both PSU are configured for 1.0 V output.

[nctnico]

Linear is the only way to go in my view.  The Ramco auction was a great shout, seen some great power supplies on there and hoping to win an Metrix AX503.

That Metrix AX503 doesn't look very compelling to me. Voltage is shown with 1 decimal only. That is pretty much useless for electronics work.

I wouldn't say that... many people just need a stable 5V or 12V with some intermediates in between, sometimes.

But you don't know whether 5.0 is actually 4.9 or 5.1 by definition. Nowadays lot's of circuitry works on much lower voltages where 0.1V makes all the difference.

[radiolistener]

I don't use 5V, almost all my circuits working from 3.3V

Microcontrollers, modules - almost all needs 3.3 V.

The main goal for 5V is to feed 3.3 V regulator.

[Datman]

Why do you need a dual supply? Do you really need high current on the negative voltage?...

If not, you could buy two different power supplies beginning from the one you need more. They could be a primary powerful one and a secondary little one for negative voltages or other, i.e. 0~15V 500mA. The secondary one could also be a dual 0~15V 500mA.

For 5V and 3.3V you could use a common USB PS, a USB cable and a little box with a simple filter (a common-mode coil salvaged from an old CFL lamp and a capacitor) and an integrated low drop 3.3V regulator.

[nctnico]

   I typically pay between $10 and $20 for older working HP supplies. If you want one with digital displays, expect to pay considerably more but they won't work any better than an old linear supply with analog meters.

I prefer digital meters. With analog meters I usually pull out a DMM to check the voltage. It might be worth checking out how accurate the meters on the Korad supply are. Recently I bought a bunch of Gophert PSUs (these are switching ones) from Aliexpress but their meters are shockingly accurate. Especially given the price.

I have HP supplies from the 1960s that still work fine. One of the nice thing about the HP supplies is that they used GOOD capacitors and I've never seen a cap fail in one.

I'd check the wiring in those. Especially when wires are soldered into boards directly. Over the years I have owned several older HP power supplies and in some I could pull the wires from the boards. Two of them needed replacing the electrolytics. In other words YMMV and I wouldn't recommend someone buying such old supplies.

[Yansi]

   I typically pay between $10 and $20 for older working HP supplies. If you want one with digital displays, expect to pay considerably more but they won't work any better than an old linear supply with analog meters.

I prefer digital meters. With analog meters I usually pull out a DMM to check the voltage. It might be worth checking out how accurate the meters on the Korad supply are. Recently I bought a bunch of Gophert PSUs (these are switching ones) from Aliexpress but their meters are shockingly accurate. Especially given the price.

I have HP supplies from the 1960s that still work fine. One of the nice thing about the HP supplies is that they used GOOD capacitors and I've never seen a cap fail in one.

I'd check the wiring in those. Especially when wires are soldered into boards directly. Over the years I have owned several older HP power supplies and in some I could pull the wires from the boards. Two of them needed replacing the electrolytics. In other words YMMV and I wouldn't recommend someone buying such old supplies.

The Korad voltage meter/preset is and was almost spot on on all PSUs I've put my hands on.  If you set i 5.000, you can bet you'd be where you need to.
The current meter is a completely differet matter, but one does not expect much accuracy there, right?

[Electro Fan]

I have found the Korad to be be consistently accurate on voltage and current.  If you ever need it there is a calibration routine that is accessible with the outside controls (no need to open the power supply).  I made some very small calibration adjustments when I purchased the PS years ago and it continues to perform consistently and reliably.  After accumulating a TEA-worthy bench full of test equipment I would have to say that the Korad KA3005P is very possibly the best value on my bench when I consider how often each piece of equipment is used along with how accurate, reliable, and enjoyable it is to use, plus what it cost.  If you don’t need the computer control feature I’m sure the same or even better could be said about the KA3005D - in fact it’s probably an even better value and arguably one of the best values among not only linear power supplies but among any entry level piece of test equipment you can buy new for about $110.

[Yansi]

But beware, there were many different versions of the very same Korad supply.  For example, at least one of my units came with Nippon Chemicon capacitors - I have no idea how that happened. Others reported cheap material PCBs n the power section - I have green standard  FR4 there.

Also, in the past, some of the early units had wrong or incorrect power transistors mounted, so the supply could easily die.

[masterx81]

I have the Korad ka3305p dual psu, spot on. Not precise at the very milliampere, but it's good enough for most uses.
Take always in mind that this type of psu have some capacitance at the output. Never "hot connect" something. Use always the on/off button.

[Yansi]

Well, you can "hot connect", but you need to be aware of what may happen.

Cranking the PSU to 30 V (or even much less) and setting the current limit to 10 mA to test a LED hot connecting it, will definitely blow shit up.

Also, using the output on/off button is a good practice anyway.

[Majorbob]

A few comments to reply to, sorry but I'm not going to add in multiple quotes.

At the moment only having 1dp isn't a massive issue and not a big concern, as someone else says I can use a separate voltmeter to get more accuracy (and I'm not adverse to adding a voltage divider if finer adjustment is required)

The PSU is going to be an xmas present (hence the limited budget).  If my usage of it increases greatly then I'll invest in a brand new supply with more programming abilities etc.

Searching ebay has shown very few PSUs available in the budget that fit my requirements at all.

Thanks for the help and suggestions, this thread will definitely be referred to if/when I'm looking at upgrading.

Edit:
I did look at separate single supplies for the same job as a dual, but with the cost and functionality of the one I've chosen I've decided going for 2 singles isn't worth the worry.

I try not to hot-connect where possible at any voltage or current.