Changing Pots to multiturn in a 30V/3A PSU

[JacquesBBB]

Hi All,

I have 30 V 3A dual bench power supply Selectronic SL-1731SB, which exists on many other names
as the Velleman PS23023, and many other ones.
See my earlier post
https://www.eevblog.com/forum/repair/tracking-problem-in-a-dual-bench-power-supply/msg529140/#msg529140
and the schematics  below.

I would like to change all the 4 front pots (6.8k  RP4,RP5,RP6,RP7) to multiturn ones  in order to get
much better precision on the tuning of the PSU.

I have several 10k Bourns 10 turns pots which  I could put there.

(Edit : In fact I made a mistake, I had only one. )

DO you see any problem that could occur from changing the 6.8 k pots to 10k ones ?

Thanks

[LaurenceW]

I'd say that you'd get away with that. The four pots are used as a mix of variable taps between different voltages (ground and Vref, for example), so a slighter higher values resitance will have little impact, there. You are still "tapping off" 50% of the same voltage, when the pot is centred, whether the pot is 6K8, 10K or even a little more.  Yes, somebody will claim that it means the source impedance offered to the circuit is higher, but I really doubt this will have a great impact. You could do worse that strap a 15K resistor between each end of the track on the variable resistor (usually labled pins 1 & 3), As this reduces the 10K pot // 15K to  around 6K. But you may even get away without those.

Try it!

Not sure I'd even BOTHER with 10 turn pots for the current limit, though. The occasions when you actually NEED to set current limits to a value whose precision you could not achieve with a single turn pot are few and far between, in practice. Remember current limitting is mainly there to save you from FRYING a circuit which developed a short fault, or something similar.

Oh, and one other thing. Ten-turn pots are a PAIN IN THE @RSE when you find that the voltage you want next is at the OTHER END of the ten turns! What you would have achieved in a single twist of the wrist now looks set to give you RSI. 2-3 turn pots might actually make for a more USEABLE instrument, and again allow the setting of the output voltage to "close enough" (for all real world apps). It's easy to get carried away with precision for the wrong reasons.

[JacquesBBB]

Thanks a lot for the advice, LaurenceW,

I will give it a try.
As it is a dual PSU, I may put  10 turns only on the  Master side.
I put 10 turns because I have them on hand. I do not have 2-3 turns pots.

For the Current, I understand your point, but  the difficulty I have at present is for very small
currents,  (and same thing for small voltage), where the the present pots are quite unusable.

[erikj]

I think you can change the pots yes, but you will likely need to change the value of some resistors as well.
R15 and R16 might need to be increased or you will get the wrong ranges.

[MrAl]

Hi,

Are you saying that the original had 6.8k and you want to use 10k?

If that is true, then the rest of the circuit will see a slight difference in series impedance, but luckily the remedy is really simple.

The fix is to parallel the new pots outer terminals with a resistor that makes the total feed through resistance equal to 6.8k again.  This works out to exactly 21.25k, but you can probably get by with 21k or maybe even 20k.

We can not do this if the resistance ratio is too large, but since yours is only 10k to 6.8k (about 70 percent difference) it will work very well.  The times when it doesnt work is when we have to go from say 10k to 1k, in which case the adjustment will turn out to be very non linear and so will not 'feel' right when the user adjusts the voltage, but for a small change the non linearity is very very slight, so it seems natural.

You are very lucky in that your power supply uses 6.8k resistors for all the adjustments.  Many of them use much lower values like 500 ohms.

One catch here though is that the new pots may be wire wound.  This may make the adjustment a little 'bumpy' anyway.  It would be best to wire ONE pot in as a test to see if you like the way it works before going through the trouble of doing all the pots and then finding out maybe you dont like the way it works.

[JacquesBBB]

Thanks MrAI,

This is somewhat what I   thought was the way to go after reading the hint of LaurenceW and other stuff on the Web.
In fact, on the web,  you find plenty of discussions on more complex settings of resistances in parallel to the
various  pins of the pot, but not much on the most obvious one which is a resistor  R3  in parallel on the whole resistor 1-3 (R= R1+R2).

From elementary Ohm's law,  I would figure that the output voltage in pin 2 is not change at all :

If V is on pin 1,  GND on pin 3, one should have  V x R2/(R1+R2) at pin 2 with or without R3.

Why do you say that there will be a change in the linearity of the pot ?

[splin]

Anyone else spot the mistake in the schematic? Transistor V18 wouldn't do much if it really was a pnp device - it's actually npn.

I wonder if (published) schematics ever have deliberate mistakes to make it a bit harder to copy? Like map makers who deliberately include errors to catch out copyright thieves.

[Alex Eisenhut]

Oh, and one other thing. Ten-turn pots are a PAIN IN THE @RSE when you find that the voltage you want next is at the OTHER END of the ten turns!

Motorize it like studio pots used to be. Maybe if you're in a hurry use a motor like this

http://www.hobbyking.com/hobbyking/store/__31677__Turnigy_AquaStar_4084_1050KV_Water_Cooled_Brushless_Motor.html

[LaurenceW]

The secret for a current limit pot is to use a standard LOG LAW (as opposed to liner law) one. This basically expands the resolution at low currents (i.e makes it easy to set 10, 20, 50 mA) while sacrificing resolution at the higher end. But then who cares if you can only set the current limit to 2.5A +/- 200mA? In practice, it most likely wont matter.

[ruffy91]

If you just parallel a resistor over the whole pot you wont change the source impedance on the wiper. Just think about it.

[nctnico]

A resistor parallel across the ends of a pot is the best way. I did that to get an odd potmeter value. I did some calculations with Excel and a resistor across the ends of a pot gives the flattest response.

I agree about the current setting. On a cheap PSU I only changed the voltage setting potmeter for a 10 turns one. For me the current adjust is just to prevent (too much) smoke.

[MrAl]

Thanks MrAI,

This is somewhat what I   thought was the way to go after reading the hint of LaurenceW and other stuff on the Web.
In fact, on the web,  you find plenty of discussions on more complex settings of resistances in parallel to the
various  pins of the pot, but not much on the most obvious one which is a resistor  R3  in parallel on the whole resistor 1-3 (R= R1+R2).

From elementary Ohm's law,  I would figure that the output voltage in pin 2 is not change at all :

If V is on pin 1,  GND on pin 3, one should have  V x R2/(R1+R2) at pin 2 with or without R3.

Why do you say that there will be a change in the linearity of the pot ?

Hi,

There are a couple ways to use a pot.

One is with the arm shorted to one end.  With this arrangement and a parallel resistor, the pot resistance no longer changes linearly with shaft rotation angle.

The other is as a voltage divider.  I think you are using this arrangement.  When used like this, the voltage output of the arm (as you say) is linear with shaft rotation, but that is only true for a driven impedance that is infinite or very high relative to the total pot resistance.  With any reasonable loading, the output voltage will not follow the rule Vs*(R2/(R1+R2)) because there is more loading effect near the center of travel than at the ends because the ends will be lower impedance than the center.  Luckily it may not be that much different than it is now though, because it must be loading a least a little anyway even with 6.8k, so with 10k there should be only a small difference.  You may not even notice it.

The time when these things really matter is when the resistance change has to be substantial, or for some kind of process control where accuracy is very important.

[JacquesBBB]

I finally did the change.
I put a 10k 10 turn pot with a 21k resistor in parallel.
I also added a 10 uF ceramic capacitor at each output.

It works perfectly. I have also the impression of a  better stability
of the voltage output with the 10k (bourns) pot.

I must say that  this is not overkill. When dealing  with  low voltage (0-5V),
the 10 turn pot is very nice to use, and I would not like less.
Going from one end to the other is quite fast as well, as the pot is very smooth, and
I put a large knob that I can turn with one finger.

I may even  do it on other places if I get hold of additional 10k 10 turn pots
(this one was scavenged from discarded equipment).
 

[OldNeurons]

I finally did the change.
I put a 10k 10 turn pot with a 21k resistor in parallel.
I also added a 10 uF ceramic capacitor at each output.

It works perfectly. I have also the impression of a  better stability
of the voltage output with the 10k (bourns) pot.

I must say that  this is not overkill. When dealing  with  low voltage (0-5V),
the 10 turn pot is very nice to use, and I would not like less.
Going from one end to the other is quite fast as well, as the pot is very smooth, and
I put a large knob that I can turn with one finger.

I may even  do it on other places if I get hold of additional 10k 10 turn pots
(this one was scavenged from discarded equipment).
 

Just another idea...
Instead of the 10 turns, why not using dual concentric potentiometers?
You will get both speed and accuracy adjustment.

http://www.bourns.com/ProAudio.aspx?name=pamm
http://www.bourns.com/data/global/pdfs/PTH90.pdf

[m100]

Just another idea...
Instead of the 10 turns, why not using dual concentric potentiometers?
You will get both speed and accuracy adjustment.

http://www.bourns.com/ProAudio.aspx?name=pamm
http://www.bourns.com/data/global/pdfs/PTH90.pdf

Aren't those actually dual gang pots where both gangs operate together using one knob, and a rotary switch on a concentric knob thus giving no independent operation of each pot for a fine adjustment?

[paulie]

Aren't those actually dual gang pots where both gangs operate together using one knob, and a rotary switch on a concentric knob thus giving no independent operation of each pot for a fine adjustment?

That's true of the first link which is probably a typo. The PDF is correct type with independent control. Much better for this app IMO. I had expensive multi-turn in my first PSU but got tired of twisting knobs so replaced with concentric. Huge improvement. Not only CHEAPER but more ergonomic and capable of better precision!

[JacquesBBB]

For me it was not an option, as I got the  10 turns pots for free  as a salvaged component
from dumpster. By the way, it was also the case for the PSU, which I had to fix a little bit.

The nice thing about salvaged equipment, is that you do not hesitate to hack it,
searching for some improvement. Right now, I am quite satisfied with the new pot,
but  I have another of these PSU (from the same dumpster) that I may try to hack in a different way,
keeping this one fully functional.

[OldNeurons]

Just another idea...
Instead of the 10 turns, why not using dual concentric potentiometers?
You will get both speed and accuracy adjustment.

http://www.bourns.com/ProAudio.aspx?name=pamm
http://www.bourns.com/data/global/pdfs/PTH90.pdf

Aren't those actually dual gang pots where both gangs operate together using one knob, and a rotary switch on a concentric knob thus giving no independent operation of each pot for a fine adjustment?

No. Each gang is controlled by a separate knob. They are totally independent.
To my knowledge, this kind of potentiometer is used mainly on certain guitars models.
The difficulty for the application above is to find adequate knobs.

[TimFox]

I have several older -hp- power supplies with concentric independent potentiometers for coarse and fine adjustment.
They can be adjusted quickly, but it is easy to go too far on the coarse, try to bring it back with fine, curse, re-adjust coarse, try to bring it back with fine, curse again, etc.  The best technique is to set fine to the center of its range, adjust coarse, then re-adjust fine and hopefully obtain the value.
A ten-turn pot is generally easier.
Note that wire-wound 1-turn pots and all 10-turn pots (except for very expensive "hybritron" units with plastic coating on the wire) have a finite resolution, and one can note the "click-click-click" as the pot is adjusted.  However, good wire-wound units are more stable than other types.

[ym58]

Hi Jacques,

I too have an SL-1731SB dual 30V/3A P/S which I am happy with but I definitely  want to change its tension pot (master, slave or both) with a multiturn pot.

I have found this rather cheap model on eBay but before I open the P/S case I was wondering if there's enough room in there to fit the quite long depth of this pot  ?

What do you think ?

[JacquesBBB]

No,

It is much too large. It will not fit.
I will post later on some more details.

[ym58]

Do you think this one would be more appropriate ?
Thks.

[JacquesBBB]

Yes this one  is similar to the one I put, maybe a little bit smaller.
So it will be OK.

But beware, you will have to enlarge the hole   for the pot handle. It is larger than the
previous one. This is what I did. I changed also the knob.
As the pot cannot be hold on the PCB, now it is hold by the front panel. See
my previous pictures.

Good luck.

[nctnico]

I wouldn't buy these kind of potmeters from Ebay. My experience so far has been that their range does not extend to zero. A 10k pot may have a minimum of 100 Ohms or so.

[JacquesBBB]

I used a Bourns 10 turn pot that goes to zero.
The one mentioned above looks like, but  if it is not certified to be a bourns pot,  it is probably not.