Would you pay $100 for a HP6115A psu ?

[BravoV]

Reading non-contiguous separated pages of scanned schematic is PITA  , decided to stitch them up into a big single page.

Sharing the HP 6114A and 6115A schematic diagram in single big sheet, stitched and applied minor post processing to sharpen and reduce the grayish background into white.

WARNING ! Huge picture !

[BravoV]

Added different colors on each boards for better clarity when trouble shooting ! 

[robrenz]

Nice job, thanks.  I did that manually with printed pages but yours is nicer

[robrenz]

Just picked up another 6115A with the 10 turn current adjust.  Very nice condition and It has the last cal tag from HP labs dated 1/25/1992 Just hooked it up to the 8846A untouched to see how far off it is. Spec is .025%+1mV.

1V +144µV = .014%
5V  -390µV = .0078%
10V -.00235V = .0235%
20V -.0025V = .0125%
50V -.008V =.016%
100V -.0179V =.0179%

Almost 22 years since its last cal and it is well within spec. 12 hour drift test after letting the reading settle for 4 hours was 9µV total span. The 4 hour initial settle drift was 400µV. Combined that is still within the 8 hour drift spec.

THESE THINGS ARE AMAZING!

[alm]

Reading non-contiguous separated pages of scanned schematic is PITA  , decided to stitch them up into a big single page.

Note that you can also get high quality scans for about $10 from ArtekManuals and Qservice. Most of the Artek ones have stitched schematics, and otherwise it only takes one email to get a fixed one. Contrast, resolution and noise is usually much better than freely available scans.

But I think I am done on power supplies. 8 (not counting the 2 BOSS bipolars) is plenty. 

Just picked up another 6115A with the 10 turn current adjust.

I think you might be the exception to the adage "you can never have too many power supplies" .

Just hooked it up to the 8846A untouched to see how far off it is. Spec is .025%+1mV.
[...]

Impressive performance, especially the low offset at the lower ranges. It desperately calls for comparison with other bench DMMs. I know some old meters from HP et al. that go for less than one eighth of the 8846A retail price .

[BravoV]

THESE THINGS ARE AMAZING!

Congratulations ! 

Can't agree more, and since I have only 5.5 digits DMM, basically when I set the meter with 3 digits mantissa, its basically won't bulge. One of the best and most stable adjustable PSU that I have ever used. 

How is the condition ?

Reading non-contiguous separated pages of scanned schematic is PITA  , decided to stitch them up into a big single page.

Note that you can also get high quality scans for about $10 from ArtekManuals and Qservice. Most of the Artek ones have stitched schematics, and otherwise it only takes one email to get a fixed one. Contrast, resolution and noise is usually much better than freely available scans.

I'm aware of that, and bought two manuals from Artek while ago, but for this HP 6114/5A full schematic, just download my stitched photo, its more than enough to see all circuit's details.

[BravoV]

Almost 22 years since its last cal and it is well within spec. 12 hour drift test after letting the reading settle for 4 hours was 9µV total span. The 4 hour initial settle drift was 400µV. Combined that is still within the 8 hour drift spec.

Looking at above numbers, just curious if this is what people called that an "aged & cooked" voltage reference and/or other precision components are much better than new one ?

[robrenz]

I think you might be the exception to the adage "you can never have too many power supplies" .

There will be a Power Designs sale in the near future.  The better HP spec especially on current and the extremely flexible remote control and series, parallel, and tracking flexibility means I am jumping ship The Power Designs are extremely good but they were not designed for accurate current sourcing.

Impressive performance, especially the low offset at the lower ranges. It desperately calls for comparison with other bench DMMs. I know some old meters from HP et al. that go for less than one eighth of the 8846A retail price .

Please don't feed or encourage the wild animals

[robrenz]

How is the condition ?

Excellent external condition, but as expected the pots need a serious clean and lube.

[GreyWoolfe]

I just started reading this thread and what a tease .  I guess I missed the boat on your August comment BravoV, the cheapest on ePay right now is $199, "used but works good.  Satisfaction guaranteed."  BUT, there is no return/exchanges.  High price is $650.  I guess the cat is already out of the bag.

Tom (with PSU envy)

[BravoV]

There will be a Power Designs sale in the near future.  The better HP spec especially on current and the extremely flexible remote control and series, parallel, and tracking flexibility means I am jumping ship The Power Designs are extremely good but they were not designed for accurate current sourcing.

Doh ... all HP 611X series lovers will be doomed, this man is jumping ship from his truck loads of PD psus, hurry !!!    ... j/k

Excellent external condition, but as expected the pots need a serious clean and lube.

Great to hear that ! 

Btw, photo shots of the process details while you clean & lube that pot please, really eager to learn & watch how you do that, must be very interesting, pretty please.

I just started reading this thread and what a tease .  I guess I missed the boat on your August comment BravoV, the cheapest on ePay right now is $199, "used but works good.  Satisfaction guaranteed."  BUT, there is no return/exchanges.  High price is $650.  I guess the cat is already out of the bag.

Tom (with PSU envy)

Tom, blame it on PedroDaGr8 and Robrenz especially after watching his restoration thread at his 6114A, I did ruin my keyboard after watching that because of excessive drooling. Both are guilty of making this infectious HP611XA addiction started here in this forum, at least to me. 

[SeanB]

Pity the PD ones will be too heavy to send to ZA unless you send then slow mail.

[saturation]

.. with this confirmation of its capabilities, I think we won't see the 6114 or 5a in the $100 range for sometime  .  Its performance is impressive  ! 

I think you might be the exception to the adage "you can never have too many power supplies" .

There will be a Power Designs sale in the near future.  The better HP spec especially on current and the extremely flexible remote control and series, parallel, and tracking flexibility means I am jumping ship The Power Designs are extremely good but they were not designed for accurate current sourcing.

Impressive performance, especially the low offset at the lower ranges. It desperately calls for comparison with other bench DMMs. I know some old meters from HP et al. that go for less than one eighth of the 8846A retail price .

Please don't feed or encourage the wild animals

[BravoV]

I think its just the matter of time that some experienced fellows here in this forum that have this gem will come up with the digital front end for it's remote terminals to tame this beast. I will be sitting nicely at the corner watching & lurking closely for this as I'm not qualified for that. 

[alm]

What would be the point? Do you often build automated test setups? Are rotary encoders and buttons somehow superior to the knobs / thumb wheels they're trying to imitate? The course voltage controls are actually digital on this supply.

Digital controls from the remote programming input will require a voltage source with a better accuracy than the power supply for it not to degrade performance. Stability would mostly be a function of the stability of the programming voltage, so this will require an extremely stable and accurate DAC. HP made a power supply programmer controlled via GPIB, but I'm pretty sure it doesn't have the resolution or accuracy for this.

[robrenz]

Agreed, but I am thinking a relay matrix that can connect the rear connections of three 6114/5As in all of the possible combinations of remote programming, remote sense, normal/auto series, normal/auto parallel, auto tracking, 1, 2, or three units, etc. with necessary connections available at a front panel including connect/disconnect from load relay would be very nice. That would be one very accurate 300VDC supply and they can float at 300V.  A simple mcu interface for all the myriad combinations of relay combinations might be appropriate here. Being able to do a few keystrokes to get any of the features above without dragging all the units out and doing a bunch of wiring on the back would be pretty nice.  Just thinking out loud, I have not thought it through yet.

The output capacitor is even connected by the rear terminals so it can be disconnected when appropriate for sensitive current limit situations.

[robrenz]

Digital controls from the remote programming input will require a voltage source with a better accuracy than the power supply for it not to degrade performance.

My new Keithley 2450 SMU  will be a >= accuracy programming voltage but only when using source read back. It will be a nice way to programmatically control the 611X's

[BravoV]

What would be the point? Do you often build automated test setups? Are rotary encoders and buttons somehow superior to the knobs / thumb wheels they're trying to imitate? The course voltage controls are actually digital on this supply.

Thats it, automated or scripted voltage & current setting from PC or mcu.

Not complaining about the encoder buttons, its just you can't jump / simulate manually with those buttons or by tuning the pot say for example from X to Y volt in short period, or a gradual but stable change like from X -> Y volt in Z sec/min/hour.

Digital controls from the remote programming input will require a voltage source with a better accuracy than the power supply for it not to degrade performance. Stability would mostly be a function of the stability of the programming voltage, so this will require an extremely stable and accurate DAC. HP made a power supply programmer controlled via GPIB, but I'm pretty sure it doesn't have the resolution or accuracy for this.

Any idea roughly on how much stability the external v-ref needed to maintain it's original performance ?

Agreed, but I am thinking a relay matrix that can connect the rear connections of three 6114/5As in all of the possible combinations of remote programming, remote sense, normal/auto series, normal/auto parallel, auto tracking, 1, 2, or three units, etc. with necessary connections available at a front panel including connect/disconnect from load relay would be very nice. That would be one very accurate 300VDC supply and they can float at 300V.  A simple mcu interface for all the myriad combinations of relay combinations might be appropriate here. Being able to do a few keystrokes to get any of the features above without dragging all the units out and doing a bunch of wiring on the back would be pretty nice.  Just thinking out loud, I have not thought it through yet.

The output capacitor is even connected by the rear terminals so it can be disconnected when appropriate for sensitive current limit situations.

+1 This too, thanks Robrenz ! 

[alm]

Not complaining about the encoder buttons, its just you can't jump / simulate manually with those buttons or by tuning the pot say for example from X to Y volt in short period, or a gradual but stable change like from X -> Y volt in Z sec/min/hour.

That's true, but for most applications you don't need the high accuracy of this power supply, however. Doing this with lower accuracy would be much easier. Then you could just hook up an isolated 10-bit DAC or so to an op-amp. Or the HP power supply programmer I already mentioned if you have GPIB. Keep in mind that this power supply may not be designed for fast slewing of its voltage or current (150ms up and 1.5s down with no load). It does not have a downprogramming circuit, for example, so lowering the voltage with a large cap connected and a light load will take forever (and may even blow pass transistors if this was as well designed as some of the Harrison ones ).

Any idea roughly on how much stability the external v-ref needed to maintain it's original performance ?

Look at the specs (0.025% + 1 mV output voltage accuracy and 0.0015% + 15 uV 8h drift). The specs specify voltage programming accuracy as accuracy of the remote source + 0.2 mV, so you would need something like 0.01% + 100 uV accuracy and 0.0005% + 100 uV drift for it not to degrade the accuracy too much.

[johnh]

Back in March I paid $40 US for a HP6114.   

Then there was the shipping back to Australia  $200  (Used a freight forwarder)
For that and some other stuff plus some things my wife bought

Some of the heat sink fins were bent. But, managed to straighten then up.

 

[BravoV]

That's true, but for most applications you don't need the high accuracy of this power supply, however. Doing this with lower accuracy would be much easier. Then you could just hook up an isolated 10-bit DAC or so to an op-amp. Or the HP power supply programmer I already mentioned if you have GPIB. Keep in mind that this power supply may not be designed for fast slewing of its voltage or current (150ms up and 1.5s down with no load). It does not have a downprogramming circuit, for example, so lowering the voltage with a large cap connected and a light load will take forever (and may even blow pass transistors if this was as well designed as some of the Harrison ones ).

Well, realistically I don't aim for the v-ref has fast slew rate more than, say 500 ms (half sec), this should be  "relatively easy" and manageable right ?

Just want to be sure, you're worry about the AC loop stability, right ? CMMIW 

Look at the specs (0.025% + 1 mV output voltage accuracy and 0.0015% + 15 uV 8h drift). The specs specify voltage programming accuracy as accuracy of the remote source + 0.2 mV, so you would need something like 0.01% + 100 uV accuracy and 0.0005% + 100 uV drift for it not to degrade the accuracy too much.

Alm, just asking as a noob, is this enough that the programmer equipped with such devices, not very ambitious though, like < 5 ppm/C voltage reference chip, and with the final driver say like using a chopper op-amp with offset like < 5 uV with drift of < 50nV/C ?

[robrenz]

The voltage programming current is a maximum of 1µA when using the required 32 kOhm resistor do you need a buffer?

[alm]

You're going to need some sort of high resolution voltage divider, most likely a good DAC. I think there's a thread on this forum about a high-end DAC based on an LT appnote. If you want to keep the floating supply, this programmer will need to be isolated from the USB connection if you use USB. I believe the 6115A has a sensitivity of 1V/V on the programming interface, so you'll need to amplify it to the 0-50 V range.

Downprogramming may be even slower than 500 ms. If there's no load, the output cap will need to be discharged through the power supply, which might take up to 1.5s according to the specs. Of course it will be much faster with a large load that quickly discharges the cap.

There's nothing wrong with using a programmer with a much lower accuracy, but it won't be a good replacement for the front panel controls and is kind of a waste of this high end power supply. If you have a GPIB interface, the HP 59501A may not be a bad deal. It's sometimes offered for $30 with few bids, and offers an isolated 10-bit DAC.

[uoficowboy]

Reading non-contiguous separated pages of scanned schematic is PITA  , decided to stitch them up into a big single page.

Note that you can also get high quality scans for about $10 from ArtekManuals and Qservice. Most of the Artek ones have stitched schematics, and otherwise it only takes one email to get a fixed one. Contrast, resolution and noise is usually much better than freely available scans.

Yep - I have the manual for the 6114A from Artek and it indeed has a fully stitched schematic. $12.50 well spent.

[BravoV]

You're going to need some sort of high resolution voltage divider, most likely a good DAC. I think there's a thread on this forum about a high-end DAC based on an LT appnote. If you want to keep the floating supply, this programmer will need to be isolated from the USB connection if you use USB. I believe the 6115A has a sensitivity of 1V/V on the programming interface, so you'll need to amplify it to the 0-50 V range.

Which thread is that ? Mind pointing that out please.

Yeah, I've been thinking about that 1V/V too, and 6115A capable of up to 100V. 

Downprogramming may be even slower than 500 ms. If there's no load, the output cap will need to be discharged through the power supply, which might take up to 1.5s according to the specs. Of course it will be much faster with a large load that quickly discharges the cap.

Looks like its not that easy isn't it ? I guess I'm a bit naive on what to expect. 

There's nothing wrong with using a programmer with a much lower accuracy, but it won't be a good replacement for the front panel controls and is kind of a waste of this high end power supply. If you have a GPIB interface, the HP 59501A may not be a bad deal. It's sometimes offered for $30 with few bids, and offers an isolated 10-bit DAC.

Still that programmer output range at +/- 10V, high voltage like at 50 or even 100V is quite challenging isn't it ?