HELP please tell me what to do ? - 6632B

[eevblogfan]

Hey

i recently recived my new 6632B and i am very happy . the mechine is a bit out (few counts )  but within 0.07%  to my 3478a , the thing who drive me nuts is the Cc mode .. i used the PS as current generator and  put 100mA thrugh my fluke 87V  and 287 ... (on the 400 mA range ) ,  and then i saw 0 current o_O . I've cheked ANY THING  and didn't saw any thing un plug .. so I shorted each multimeter ... it was my 87v's fuse ,  by mistake i killed my 287 fuse ass well ... I cant belive it !

Can someone tell me . can i make the 6632B's CC faster ?  thank you

[AndyC_772]

Have a look at the later posts (#9 onwards) here - I've made some measurements on one of my own 6632B's which you might like to repeat in order to see if yours is behaving as it should.

[free_electron]

What did you do. . Set current lim , turn output on,  then plug in banana ?  Most likely. You dumped the output cap energy in the fluke: fuse popped.

The correct procedure is:
Output off.
Set current lim and voltage.
Connect everything.
Turn output on.


The power supply will now ramp the current from zero under cpu control. There will be no overshoot .

When done : turn output off.

The on/off i. These machines is not a clunk-relay. It is dac driven.

[eevblogfan]

The bananas where in ... i only switched on the output and saw 0 current flow - stupide annoing inncident !


Sorry my english  (i am posting viah cell phone )

[Slowflyer]

The 6632B is designed to mainly be a CV supply. Current limit is slow to kick in, and the current can be many times higher for tens of milliseconds. The workaround is to short the output, set current limit and turn the output on.
After that, you can connect the load and remove short. It's PITA though.

[eevblogfan]

hey

I am awear it's mainly CV supply . But can I do something within the circutry in order to drop that into few milisecoands or perhaps 50uS or such >? ( I really liked the speed of my 3610A )

thank you in advance !

[AndyC_772]

I determined in my own measurements that:

"If the load resistance is too low for CV mode - ie. if the current would be above the prescribed limit - then the supply will deliver about 20mC of additional charge into the load before switching to CC mode. This is fairly constant and predictable; it'll deliver an extra 200mA for 100ms, or an extra 3A for about 6ms. The product of [(extra current over and above the programmed limit) * time] remains remarkably constant."

I'm not aware of a way to defeat this behaviour. If you have a specific need to do so, then you might need to build an external current limiting circuit to sit between the 6632B and your load.

Bear in mind that many (most?) circuits that are designed to operate off a constant voltage will draw a significant amount of current at start-up in order to charge their bulk storage capacitors. If they don't get this current, because a PSU has gone into CC mode or their supply cables are too resistive or inductive, then they can fail to start up properly. In some cases this can even lead to oscillations that end in damage.

[Slowflyer]

It is either impossible or very hard, since the regulation is controlled by software using DAC's and ADC's instead of the pure analog way. The software should be rewritten.

[eevblogfan]

hey

my need is simple - CV _ 50uS_> CV .
 
no need for any extra time of delay -_-

[bingo600]

Did you check the servicenotes
http://litstation.marketing.agilent.com/litapp/SearchSN.do?method=openExternalSNSearch&prodNum=6632B

I did a new of them on my 2 units.
But then i haven't measured current on an external meter.

/Bingo

[eevblogfan]

hey Bingo

I have no idea from where I shall be getting that rom or how to treat him , sorry for being such newbie 

please tell me - have I got acces to the proggremed current ? ( say I put 0.1A on the keypad . the DAC out puts say 100mV somewhere , can you spot me where ?

My idea is to build some CC externaly and the CC will be controlled via the DAC . is that possible ?

thank you in advance !

[mzzj]

It is either impossible or very hard, since the regulation is controlled by software using DAC's and ADC's instead of the pure analog way. The software should be rewritten.

If you take a look at the schematics there is no software or digital parts involved  in the control loop. Nonetheless its not going to be  easy to modify and maintain stability under varying load conditions.

BTW, There is second "hard" limit on 6632B that acts without delay and it kicks in above 6 amps.

[mzzj]

What did you do. . Set current lim , turn output on,  then plug in banana ?  Most likely. You dumped the output cap energy in the fluke: fuse popped.

The correct procedure is:
Output off.
Set current lim and voltage.
Connect everything.
Turn output on.


The power supply will now ramp the current from zero under cpu control. There will be no overshoot .

When done : turn output off.

The on/off i. These machines is not a clunk-relay. It is dac driven.

Does not work, 6632B will start under voltage control and current control loop is too slow to react.  AndyC_772's numbers of delay sound about right.

[bingo600]

I determined in my own measurements that:

"If the load resistance is too low for CV mode - ie. if the current would be above the prescribed limit - then the supply will deliver about 20mC of additional charge into the load before switching to CC mode. This is fairly constant and predictable; it'll deliver an extra 200mA for 100ms, or an extra 3A for about 6ms. The product of [(extra current over and above the programmed limit) * time] remains remarkably constant."

Would that correspond to the energy in the "Output Caps" by any chance ?

I have 2 different PSU's one labeled HP and one labeled Agilent.
They have different firmware rev's , and  definately have different control loops.
It can be seen in the profiling i made here https://www.eevblog.com/forum/buysellwanted/hp-6632b-power-supply-electronic-load-in-uk/msg191834/#msg191834

They're 0..20v in 100mv steps i think it was , with a 33R load.
And the deviation from the gpib "set" to the measured 34401A and reported 6632B values.
The newer Agilent ramps up like a staircase ,and the older HP just adjusts.

/Bingo

[mzzj]

I determined in my own measurements that:

"If the load resistance is too low for CV mode - ie. if the current would be above the prescribed limit - then the supply will deliver about 20mC of additional charge into the load before switching to CC mode. This is fairly constant and predictable; it'll deliver an extra 200mA for 100ms, or an extra 3A for about 6ms. The product of [(extra current over and above the programmed limit) * time] remains remarkably constant."

Would that correspond to the energy in the "Output Caps" by any chance ?


/Bingo

Nope, its clearly not related to output cap(s).

[eevblogfan]

hey

I've found the CC "prog" voltage is okay - Ie , when entering " 2A" to the front panel , when pressing enter , the programmed current is updated straight away , so ther's nothing relating to to digital circuit , it's all analog ,

Does any one know where shall I be looking for ?  I suspect some 1uF or so capacitor some where in the CC loop , can someone help me ?

[eevblogfan]

hey
http://cp.literature.agilent.com/litweb/pdf/5962-8119.pdf

page 79 , can you see something that I miss ? , I didn't found any 0.1uF or so cap there ! , can you find ?

thank you in advance !

[eevblogfan]

 :anyone:

[Chris.M]

If the 0.1uF cap you are looking for is C365,
then its on page 79, in the voltage control ckt section.

It is 45mm from the left and 55mm from the top of page 79.
Above NORM N/A FAST.

[AndyC_772]

Are you looking to modify the 6632B to speed up its transition from CV to CC mode, or fit an external current limiter?

It looks as though CC_PROG represents the set current limit, and if you're looking to build an external limiter, isn't that all you need to access?

If you'd rather modify the 6632B itself, then I'd start probing around with a scope to see what exactly happens when. I don't see an obvious time delay in the comparator circuit that generates CC_DETECT+ (or CCN_DETECT+), but that's not to say there isn't one that I've overlooked. Another possibility is that IMON_H is filtered, so have a probe around the low and high range current monitor amps (page 78) and see what their response looks like when you apply a step increase in the current drawn from the unit.

On the low range I see a net called VMON passing through three 1MOhm resistors in series (R488,489,490), so that wouldn't take much capacitance to form the R-C filter you're looking for. On the high range, I'd start by looking with a scope at the output of U309 and go from there.

[eevblogfan]

hey

thank you AndyC_772 and Chris.M

Chris.M : isn't that cap belongs to the CV ? , if so , I shouldn't touch it , don't you think so >?

AndyC_772 : I am going to probe but I am not quiet sure what to expect ,

I am willing to modifie the 6632B , the external CC is my last option in mind

looks like the CC is working with positive feedback ( faster - isn't it ? , oh and more stable at 5A and above ) 

is there any 6632B owner who might help me with that problem  ?

tomorrow I am in the job , perhaps all week , But I have got friday fre . so perhaps I'll probe the supply only then

Cheers !