DIY bench power supply PSL-3604

[Liv]

There are manual, but to translate it into English - very hard work for me. So I just translated table of contents. Interesting paragraphs from the Adobe PDF Reader can be copied and pasted into the Google Translate.

Attach a two-volume archive, as the size of the document 1.4 MB.
First, extract PSL-3604_v2e.z01 from PSL-3604_v2e_z01.zip and unzip PSL-3604_v2e.zip.
(ext .z01 not allowed this forum)

[SimonD]

Hi,
Any progress with your electronic load project LD-8020 ?
Looks very interesting and impressive !

[Liv]

Progress is weak, with little free time. But gradually do.

[Warhawk]

Very nice and professional project Liv.
Do you consider the data here as final ? Are there some bugfixes ?

[Liv]

This is the final version. Circuits and boards have version 3 with all patches. Perhaps in the future there will be a new firmware with some additions. Current firmware version - 2.02.

Firmware update: version 2.03.

[SimonD]

Hi LiV,
So after one year later is there any progress with your very promissing LD-8020 electronic load project ?

[Liv]

Currently, only the layout of front panel board is ready.

[SimonD]

Great! This is a step forward ! 

Any progress in power stage ? Can you give us some informations about your experiments ? (Stability, thermal behavior etc.)
Maybe some detailed schematics ?
I 'm working in a similar project this time, and any collected experience  from others will be useful and appreciated !

Thanks.
Simon.

[Liv]

The power stage in the design phase. I am now researching the scheme of Agilent N3304A.

[SimonD]

Thanks for your quick response
Excellent work. Very professional !!!!!

[Electronix]

Hi liv,

really great work:) it is a very nice power supply.
I had only one short question, over sw settings I saw it is possible to switch off the down programmer.
How did you realize that ? Cause I see no switched signal in the down programmer.

Thanks

Sven

[Liv]

Thank you! When the output is enabled, the DP is always active. When the output is off, the DP can be enabled (DAC = 0, ON = 1) or disabled (DAC = 0, ON = 0). Of course, it would be better to have a separate DP control, but here it is not provided.

[Electronix]

Hi liv,

i have a sho question to your current limiter which is implemented with the components around vt9.
Will that work ? I ask me the question by myself cause over the current sensing resistors r27,r28,r25 and r26 it will not reach the 0,7 volts base emitter voltage at for example max 6 amperes or do I'm going wrong ?i try now to solve that via adding a additional 15k resistor between collector and base. That should work than.

Cheers

Sven

[Liv]

Current protection on VT9 used in special cases and in normal operation the PSU never works. For example, it protects the output stage in case of short circuit the OUT+ terminal to the GND (not OUT-) when adjusting or repairing the device. Protection threshold set by a BE voltage (about 0.7V) and R25-R28 resistors values. This threshold should be higher than the guaranteed maximum PSU output current.

[Electronix]

Hi liv,

thanks for the answer. I see now a little bit clearer.
I have now build one prototype with a similar analog stage as yours. Only that I use
one voltage which comes from a switching power supply and is pre regulated 1,2 volts higher than the output voltage.
The control stage with current mirrors and op amps is the same as yours only I use max 44246 as op amp and op27g for voltage and current amplifiers.
I have the problem that I cannot go to 0 volts. My design goal is to adjust voltage in 1mv steps.
I don't know if that will work with the design.
I can only reach 0,3 - 0,4 volts as a minimum.
What I have not implemented is the higher ground which you have realize with op amp (com).
Could that be the problem?

Thanks sven

[Liv]

I have now build one prototype with a similar analog stage as yours. Only that I use one voltage which comes from a switching power supply and is pre regulated 1,2 volts higher than the output voltage.

Using a switching preregulator has two drawbacks: a high level of high-frequency noise and poor PSU dynamic when adjusted upwards due to the need to charge the capacitor at the regulator output.

I have the problem that I cannot go to 0 volts. My design goal is to adjust voltage in 1mv steps.

If you do a precision power supply, it must be purely linear. Or use low-noise preregulator working at mains frequency. An example of such amateur PSU with 1 mV step here.

I don't know if that will work with the design.
I can only reach 0,3 - 0,4 volts as a minimum.

The PSL-3604 does not have such problems. There are two conditions to ensure zero output voltage: opamps supply voltage must be bipolar and unipolar DAC and ADC scale should be shifted up to a few hundred millivolts.

[Electronix]

Hi liv,

thanks for your quick reply.
I use for all op amps +- 15 volts.
I try to find out what the problem is.

Cheers

Sven

[Liv]

The control voltage to set the voltage and current to be formed with respect to a level of 100 - 300 mV. The PSL-3604 is a signal COM, which is obtained from the reference voltage using resistor divider. Because unipolar DAC can not provide output voltage near zero.

[Electronix]

Hi liv,

for testing I use at the moment a analog pot which is connected between ground and 5volts.
I not test it with the Dac at the moment.

[prasimix]

How close to zero are you now when connect control input to the ground?

EDIT: sorry didn't see some of your previous post. Anyway I go down to zero even without using Liv's trick with shifting COM level to few hundreds mV.

[Liv]

for testing I use at the moment a analog pot which is connected between ground and 5volts.
I not test it with the Dac at the moment.

Opams have a offset voltage. Therefore, accurate PSU output voltage can be achieved only after digital calibration. To compensate for both negative and positive offset with unipolar DAC, control signals must be level shifted by few hundred millivolts. Pots similar unipolar DAC, they can not provide the control voltage below zero.

Anyway I go down to zero even without using Liv's trick with shifting COM level to few hundreds mV.

This is not my trick, a similar solution is used, for example, Agilent U8002. If the offset is not done, the digital calibration of the zero point is impossible. The error at zero is dependent on the accuracy of the used opamps.

[milkman]

Did anyone make BOM for this bench psu. I opened attachments in posts but i couldnt find it?

[Liv]

Bom:

[jbb]

Hi Liv

Thanks for posting this - it looks like a good design and I want to try out a modified design.  I really like the output stage design; I think the the automatic range switching and missing auxiliary supply protection (using U3) are very nice.  So much better than a tap changing relay!

There are some aspects of the design of the Power Supply Board that I would like to know more about.

• In the regulator control board, VT9 and VT10 obviously pull current out of the driver stage to adjust the output voltage.  I can see why they are used in this way, but can you explain why their bases are connected to the voltage divider (R73+R74) / R77?  What advantage does this offer compared to connecting to ground?
• What protection does the Remote Sense Protection circuit (VT11, VT12, VT13 etc.) offer?
• What do C28, C29, R42 and R54 do?  Are they for damping of output voltage transients?
• What do C51 and R98 do?  Are they for damping of CC  to CV transition?

Cheers
jbb

[Liv]

So much better than a tap changing relay!

Multi-level output stage provides a good dynamics in the full output voltage range. Using relay takes time to switch on the relay and charging filter capacitance. When the complexity of the load when the PSU must sharply raise the voltage during CV-CC transition, power failures are possible. For external programming via USB, this PSU is able to change the voltage up quickly, which may be important in automatic measurements. PSU with switching preregulator has similar drawbacks as the PSU with the relay. Additionally added interference. At the same time, the PSU with multilevel output stage also has disadvantages: the increased complexity, the relatively low efficiency. This solution is good if, when you need a low-noise power supply with good dynamics and average efficiency.

In the regulator control board, VT9 and VT10 obviously pull current out of the driver stage to adjust the output voltage.  I can see why they are used in this way, but can you explain why their bases are connected to the voltage divider (R73+R74) / R77?  What advantage does this offer compared to connecting to ground?

This divider forms a local negative feedback. It is good for the dynamics and stability of power supply.

What protection does the Remote Sense Protection circuit (VT11, VT12, VT13 etc.) offer?

This scheme is similar to the Agilent 6622. This limits the output voltage if the SENSE inputs are very different from the output.

What do C28, C29, R42 and R54 do?  Are they for damping of output voltage transients?

These chains form the PSU frequency correction. This is a very delicate question. If we use a differential amplifier to produce a voltage feedback signal, then we have no right to connect the frequency correction circuit to the PSU output. It can only be connected to the output of the differential amplifier. The required depth of the correction can be achieved only in the case of inverting error amplifier (see. Agilent scheme). But this complicates the PSU. So I compromised - has set up two symmetrical frequency correction chains, bypassing the differential amplifier. In general, all questions about frequency correction is better to look at PSpice, where I spent a lot of time.

What do C51 and R98 do?  Are they for damping of CC  to CV transition?

This chain was added later. It eliminates the overvoltage in the transition between the two levels of current in CC mode. Initially, in the simulation, I missed that. The action of the chain shown in the figures below.
[/list]