BWD-603A Bench Supply

[Clay1905]

Hi Folks,

Almost two years ago I bought this power supply for not much more than a song. At the time it was about all it was worth too. Very little worked. Gradually I have been able to get some sections to come back to life. However, the stage I'm presently working on is rather less co-operative.

This stage is a sort of three in one set up. Using the same op-amp chip, this circuit can be either:
A bipolar supply (-15 to +15 VDC),
A power amplifier (Av=10),
or a fixed value 5VDC supply.

In every case, if a positive output is called for, the output is going to be +3VDC. This is regardless of which configuration is switched in. Or to put it another way, regardless of whether the op-amp is inverting or not.

Attached are two schematics.
The first is titled 15V Fixed. This supplies those voltages the circuit in question requires to function.
The second is called PowerAmp, and is the circuit in question.

There are a few errors within the schematic, the only one that isn't just a labeling typo is the -20V feeding D208 (top right). This is actually the same -25VDC as seen feeding R215 (lower right).

Rightio, some detailed info: (Pin 6 is the output pin of U200)

Bipolar Power Supply:

RV200 centered. Output= 0VDC Pin 6= 0

RV200 fully CCW. Output= -15VDC. Pin 6= -1.5VDC

RV200 fully CW. Output= +3VDC. Pin 6= +13.8VDC

Power Amplifier:

Input= -1.5VDC output= +3VDC. Pin 6 = +13.8VDC

Input= +1.5VDC Output= -15VDC Pin 6 = -1.6VDC

Fixed 5 Volt Supply:

Output = 3VDC

U200 Pin 2= +3.8VDC Pin 3= +4.8VDC Pin 6= +13.8

As my theory doesn't extend far enough, I can't see what is happening here. I believe that because both inverting and non-inverting inputs produce the same 3VDC, the fault is not with the LM318.
Exhaustive testing has revealed some failed components, and subsequently, misplaced connections that caused the failures. The wires are back where they belong, and the dud parts replaced.
I believe I can confidently say that all the components are working just fine at present.

I think this is all I can add at the moment.

Regards,
Clay.

[Clay1905]

And now,

The schematics.

(Who's a little vege-head then?)

Clay.

[trobbins]

Can the LM318 be removed (is it in a socket)?

If so then does the power amp stage idle with ok DC voltages such that bias current levels are ok and the output voltage is close to 0V? And are appropriate voltages presented to pins 2 and 3 of the LM318 socket, for the various switch settings?

[Clay1905]

Originally Posted by trobbins

Can the LM318 be removed (is it in a socket)?

If so then does the power amp stage idle with ok DC voltages such that bias current levels are ok and the output voltage is close to 0V? And are appropriate voltages presented to pins 2 and 3 of the LM318 socket, for the various switch settings?

Thanks for your reply.

I must confess to two things. Firstly is the fault has become known. Secondly, it was such an obvious thing I was completely blind to it.
R211 has gone extremely high (open for practical purposes) and so Q202 was not outputting sufficient to make the feedback loop work. This is why the 13 odd volts on U200's output. It was flogging itself trying to get more voltage onto pin 2 via R204, but it just wasn't happening.

Thanks again,
Clay.

[trobbins]

Good to hear another BWD is operational again. I have a general purpose BWD 255 bench supply from 1970'ish - it was working fine, but for peace of mind it needed some safety upgrading around the mains wiring and earthing, and I replaced the e-caps.

[Radio Wrangler]

D203, D204 D205 D206 D209 Form fairly crude current limiters for both the positive and negative-going output sections.

If the supply has been left running for a while into a short circuit, or excessive load, trying to drive a positive voltge into it, D203, D204, D209 will start conducting and clamp the base voltage on Q202 with respect to the +25v supply. This clamps the voltage across R214 to about 1.4v. Ohms law says this will limit the current to 2.8 Amps.... these are rough figures.

Meanwhile, R211, R207, R208 get almost 25v across them, less the 2.1v in the three diodes. Q200 will be saturated hard on.

23v, 47+10+47 Ohms and we get 221mA in 47 Ohms, that gives 2.294W

So the 1W rating of R208 is somewhat inadequate, and R211 gets the same amount of power. R211 isn't marked as being a high power part, but it needs to be 3W like R208.

R207 gets 500mW so it needs a suitable part.

Remember the +25v supply doesn't go through the regulator chips so it doesn't get the benefit of their current and thermal limiters.

The comparable resistors on the negative going section also need beefing up. With only two clam diodes in the string on this side, the current limit is set to half that of the positive side, so resistor dissipations in current limiting need only be a quarter of the rating of their positive counterparts. They are somewhat less overrun if current limiting.

Judging by the jump in component numbering, D209 looks like a later addition to boost the current available when supplying a positive output voltage. It looks like whoever made the change didn't redo the resistor power dissipation calculations. No surprise R211 got burned out.

Damage to R211 was the clue that sent me down this trail, and ended up finding a design whoopsie.

David

[Clay1905]

Now there's some interesting and useful information. Thank you for taking the time to work this out.

There are a few little errors apparent even to a rank amateur like me. Primarily typos and schematic errors. Components mis-labeled on the silk screening etc.
I have wondered about this unit's history too. Being a "lab supply" I wonder what experiments it has been subjected to.

As I mentioned above, there were crossed wires, and subsequent damage. Now I see that R211 met it's fate when somebody plugged the output wire from the power amplifier onto a ground pin. This was how I received the unit. There are times when carefully taking pictures so everything goes back where it was aren't all that good an idea.

A private correspondent has pointed out that Q200 might have gotten a bit over-excited too. It has. They suggested that I test R206 for potential difference across it with the bipolar amplifier outputting a negative voltage. At this time Q200 should be switched off and R206 not conducting. It is.

I would consider myself a bit of a learn-as-you-go'er. This power supply is proving to be an excellent lesson. I have quite a bit to digest. Then it's onwards with the signal generating section.

Clay.

[Radio Wrangler]

Notice that in amplifier mode, the feedback is shunt type and the voltage gain is inverting. This configuration without an input buffer means that the overall gain is a function of the source impedance of whatever is providing the signal.

David