Power supplies - any reason not to buy "vintage" HP etc.?

[aeberbach]

I've just watched the video.

There is an abundance of gear available second hand, some of it once-premium gear like HP/Agilent E3630A that could be had for around half the price of a new DP832. Is there any good reason not to go vintage? The Rigol has better specs and maybe better accuracy but are there any compelling features to make a hobbyist want it over a "traditional dumbarse bench power supply"?

[gilbenl]

FWIW-I'm fairly new to the game and my main PS are an Agilent E3610A and a 1965 era HP/Harrison 6113A. I picked up the agilent for $75 and its an outstanding unit. For $125, you should be able to get one that is virtually brand new-don't settle for less. There are slightly different model numbers for different ranges-the 3610 is 0-15v, 3A. The other, lesser known models often go for less.

The 6113A is my favorite. These are hit or miss on ebay, but disgustingly simple to work on if you get a dud. I got mine for $35 and had to replace a cap, add a cal resistor and clean the thumbwheels. The abs difference between min and max at a setting of 2.5V over the last 48 hours is 0.00043V.

I would recommend a dual output PS if you can swing it.

[radiomog]

get something with current control

dual output is nice.

[lowimpedance]

 Of course it would depend on your location as power supplies would be generally heavy and shipping would need to be considered as well.
However if you can pick up an oldie for a 'good price' then absolutely. Usually there is enough service information available and most are easy enough to repair.
Don't ignore other brands for just HP either. And as stated try and get at least a dual output.

[Electro Fan]

I like vintage gear a lot but one feature that is kind of hard to find on vintage power supplies is an output switch (separate from the main AC switch).  I guess back in the day and maybe even today user's didn't/don't like/want/need them?  Maybe they aren't so useful?  On the other hand if you look at most new power supplies today in the $400 and up range (and sometimes less) they come with output switches - so maybe they are useful?  Personally, I kind of like having an output switch on a power supply.  YMMV

[c4757p]

Most well-designed, older, linear power supplies have decent startup behavior, so you can just use the mains switch as an output switch (this is, of course, not as useful for multiple-output supplies...)

Obviously, test first. Some overshoot. HP and Power Designs seem to be good - I don't have much experience with others.


That said - I'm hugely in favor of buying used power supplies. They (or at least, the ones that lasted to 2015) were just so well built. I've bought quite a few and rarely had to do any work on them. I've done some preemptive capacitor swaps, and I've replaced dead indicator lamps, and that's about it. However, you don't get nice new features like PC control.

[TimNJ]

I think old HP (or similar) gear is great. My two lab supplies are an HP 6214A and an HP 6216A. I got both off eBay for around $40/ea I remember. One needed new capacitors and the other I replaced the caps preemptively. They're not particularly high power. 10W I think? And the plastic is very brittle, but for $40, you can't beat it, I don't think.

[VK5RC]

+1  for older HP psu,  well made,  no unobtanium ASICs,  lots of documentation and repair experience,  also very quiet electrically (bit noisy physically).  I have had a good run with 66xx series psu.

[MadTux]

I'd rather buy an old HP PSU than any new china PSU. Even the new HP PSUs aren't that good any more, as seen here:
https://www.eevblog.com/forum/testgear/video-teardown-repair-and-analysis-of-an-agilent-e3642a-dc-power-supply/

I like the lower E361x series however. Nice 10 turn pots and no custom microcontroller on the display PCB, only a MAX 1367 (or some other 3.5 digit DVM ADC) and some logic .

I would look for units with good precision pots, clean internal design and no custom parts. If it craps out in 5 years, you can repair it. If a Rigol 832 craps out in 5 years and there is a fault in the fancy display, you have a problem.

 Also depends on what you're doing. If you're working with high voltage or high power analog electronics (like inverters and stuff) you certainly don't want a fancy PSU like Rigol 832, since you'll fry the display electronics pretty soon.

[nanofrog]

Personally, I'm a big fan of used PSU's.  You get a lot more value for your money, especially if you opt for dual or triple output units. They also happen to be reliable, and they're usually repairable if they do break (manuals even tend to include schematics). Far better alternative than the entry level stuff out of China IMHO.

As a brand, it's hard to go wrong with HP or Power Designs. But there are also other names out there that are a bit less known or sought out by collectors, so can be had for less. Brands such as BK Precision, Systron-Donner, Lambda, and Kepco for example.

Not sure what you're actually looking for, as there are a lot of used alternatives from analog to digital input models.

[Electro Fan]

I think old HP (or similar) gear is great. My two lab supplies are an HP 6214A and an HP 6216A. I got both off eBay for around $40/ea I remember. One needed new capacitors and the other I replaced the caps preemptively. They're not particularly high power. 10W I think? And the plastic is very brittle, but for $40, you can't beat it, I don't think.

I have a 6289A (40V 1.5A) and a 6285A (20V 5A).  Both were damaged pretty badly by GE tech who had attempted to repair them.  They were free and time to repair them was minimal.   Also have an 59501B DAC that I use to control the 6285A with the PC.  Had them for several years now.   Agree with above post.  Not bad general purpose supplies.

I have made many power supplies over the years to get by.   The very first thing I ever built using a micro was this power supply.   VF display.  Keypad was from a girl friends dead phone.    It has RS232 that allows you to program and read back.    Also has a separate set of banana jacks for a volt meter function.    Uses a Motorola MC6801 to date it (early 80s).   I still use it.

Keypad was from a girl friends dead phone.   I guess that's better than if her phone was still working.
- there are an endless number of ideas and lessons learned to be gained in this forum 

[anotherlin]

Go ahead for old PSU, like HP or even Harrison.
Linear power supplies are working with same principles.
In fact, I've the feeling that the technology reached full maturity in the middle of the 60s.
So I don't think the newer Rigol or even newer Agilent Keysight brings anything new on the table.

Unless you need accurate programmability and/or measurement using GPIB or USB, there is no need.
An old PSU can do the job for 1/5 of the price.
Plus if it fails, the service manuals are readily available and they are made with "standard" components, easy to buy replacement, and through hole makes it easy to repair.

 

[nctnico]

I like vintage gear a lot but one feature that is kind of hard to find on vintage power supplies is an output switch (separate from the main AC switch).   Personally, I kind of like having an output switch on a power supply.  YMMV

Same here. I usually add an output switch myself if a PSU doesn't have one.

[theatrus]

Harder to find, but I've had good luck with the Xantrek/Sorensen single output units. Real multi-turn pots and a no frills interface.

[rx8pilot]

I have a pair of HP/Agisight 6653a's and would not trade them for any of the new options on the market. They were cheap on ebay.

Ultra clean, bulletproof, programmable, GPIB control. Other bits on my bench are high tech and new, but the HP psu's are fantastic.

[nctnico]

I like HP's switching PSUs. They are light to move and still pack a punch. By adding a temperature controller board for the fan they are silent as well.

[saturation]

To the OP, none so long as it meets your criteria for what you need from a PSU.   Some considerations:

Second hand market could be non-existent in some countries.  In the USA, seconds from high end labs are abundant, and you have so much choice of research grade lab supplies and shipping is generally cheap.

Is is repairable, old supplies invariably have some faults somewhere either on receipt or the future.   I only buy units with full schematics, which you can download and examine before you buy.

Much electronics in general during the 1960s-1980 were built during the peak of the space race, so aerospace and military consumers, as well as potential subcontractors [ meaning almost the entire T&M market] were always in mind, so durability and high performance was the norm. 

[barnacle2k]

IMO the HP / Agilents are way overpriced even on the used market. (At least here in Europe)
Price / performance ratio for "new from china" is much higher, and the performance figures are close.
If you don't have a specific need for the performance advantages of those supplies.
And for the 30years+ old units i would be cautious if they are still in their specs. (don't assume, measure)

btw:
I don't see many use cases for a programmable PSU outside automated test systems.
They seem cumbersome to use in the lab, lots of button presses and menus to navigate just to set voltage and current. 

[Yago]

IMO the HP / Agilents are way overpriced even on the used market. (At least here in Europe)
Price / performance ratio for "new from china" is much higher, and the performance figures are close.
If you don't have a specific need for the performance advantages of those supplies.
And for the 30years+ old units i would be cautious if they are still in their specs. (don't assume, measure)

btw:
I don't see many use cases for a programmable PSU outside automated test systems.
They seem cumbersome to use in the lab, lots of button presses and menus to navigate just to set voltage and current. 

Dunno about the programmable bit.
Hit volts, 5, enter
Hit current .05 enter
Hit output on

Not so much fuss for me, can be a little round the houses when raising the volts incrementally, but give and take when shopping on a budget.

Prices in EU, it is a matter of time and timing.
Again KJDS here is your man, 6632b for ~£140, so good I bought two!

[Alex Eisenhut]

Well, I have a HP6284A which I bought for 1$ at a former employer. It was crispy inside and I repaired it.

1) It is quite physically loooooooong with a giant resistor just waiting to stink up the place when you draw a lot of current near the limit.
2) It has some weird turn on behavior, like going negative then overshooting a bit before settling, especially when set to near 0V output. It's all over in under 2ms but you might want to scope that behavior. Turn off is also eventful, hitting -500 to -600mV rapidly and taking 2-3 seconds to decay to 0V.
3) It uses some weird old clunky wirewound trimmers inside and if they go bad it might be hard to find replacements.
4) It goes out of alignment quite quickly, like within 6 months you need to re-adjust the front panel readout.
5) The bezel breaks easily.
6) The meter switch can go wonky.

But overall it's sturdy and easy to understand, and you can try to use the analog programming features if you need em. Manual is available online.

[JuiceKing]

I've just watched the video.

There is an abundance of gear available second hand, some of it once-premium gear like HP/Agilent E3630A that could be had for around half the price of a new DP832. Is there any good reason not to go vintage? The Rigol has better specs and maybe better accuracy but are there any compelling features to make a hobbyist want it over a "traditional dumbarse bench power supply"?

The E3630A and E3610A are compact, stable and quiet supplies. No fan...no noise. There's no current control on the E3630A, and single-turn pots make it hard to adjust voltage as precisely as you can with the 10-turn pot on the E3610A. Also, there is only one display so you have to push buttons to view the different output voltages. But I find that when I do need the E3630A, I don't adjust voltages much so it doesn't matter. The two together cover a lot of ground. I found prices to vary widely on eBay. Mine were real beaters but were easy to fix (with somewhat pricey replacement parts from Agilent).

[Jidis]

I had a somewhat different experience going with an old HP 6205C last year, but I think most of it is probably a rare case due to someone's improper clean/lube work (switches were gunked up). The thread I posted in is here if anybody wants to look at it. There are some pictures of the disassembled pushbuttons in there for curious types.

It has been working fine since then, but having read some of the descriptions of the setup procedures for specific applications or current limiting, I have to wonder if I'd been better off with one of the newer import supplies where I could just press a couple buttons and be done.

George

[Rupunzell]

Never really like hp-Harrison lab power supplies. They were a mixed bag of good to ok. Others like Kepco, Power Designs, Lambda & other similar tend to cost less (brand recognition) and offer similar to better performance -vs- cost. There is little to zero wrong to purchase older "vintage" lab-bench power supplies as they are often well built, quite repairable and built to last unlike many modern plastic wonder techno products that are intended to be tossed if dead or when their specialized internal bits are NLA.

Digital meters are good, do asks the question, does the DUT really need that degree of power readout precision? Essentially, don't discount the actual usefulness of a vintage power supply simply due to it's wiggle meters.

As a rule, lab bench power supplies must be darn near un-bustable, live after being severely abused and continue to function with absolute reliability. That said, bench brewed power supplies can meet testing needs better than bench supplies at times, specially when an entire system is being built and tested.

Question is, does a circuit really need to have ultra precise regulation, ultra low noise, ultra low impedance from DC to daylight  to function properly or should the circuit in question function properly within the system power it will live in which can be quite different than being powered by a lab/bench power supply.

Often times, a lab-bench brew power supply that has several fixed outputs can be more useful than a single or dual lab-bench supply.


Stuff to consider.

Bernice