Oltronix LABPAC B60-1T switch probelm

[adinsen]

I got myself a beautiful Oltronix LABPAC B60-1T power supply yesterday. I bought it with the known issue that it cannot supply the rated current. My measuremens confirms that.

This is a 1970's precision linear lab power supply with three ranges: 0-6V 4A, 0-30V 2A, and 0-60V 1A. It's been discussed before here and here.

The PSU is very much 1970's with a big decked four way rotary switch to change between the ranges, 10 turn precision potentiometer to adjust voltage, a beautiful hand drawn main pcb with a temperature compensated zener diode, BC107's, and dual 2N3442 high voltage power transistors in the output.

During my testing, I noticed that in the 0-6V range, the PSU could supply much more than 6V. The 6V limit is implemented to protect TTL ic's (according to the manual) and is only activated in this particular range. It consists of a 2N3055 effectively short-circuiting the output when the voltage exceeds 6.7V causing the current limiter to kick in. The 2N3055 was blown.

I thought there must be a reason for this and during my search I found another issue namely that a 5 Ohm discharge-resistor supposed to discharge the main filter caps when 'output-off' is selected does not kick in when it should. It's there to discharge the filter caps when output-off is selected which is between 0-30V and 0-6V ranges on the switch. The switch not only changes the range on the linear regulator, but also changes windings on the transformer. The particular deck is buried well, but I think I can see that the finger supposed to touch the rotor in the off setting has melted away. I tried taking a picture, but I realize I have to disassemble the switch to investigate it further.

I've attached a schematic of the PSU for those interested.

I'm very much interested in tips on repairing decked swtiches like this! I could replace it with a relay board, but would prefer to get it working as I like the originality of the design

[adinsen]

Just a quick update as I removed the nuts on the ends and pulled off the outer deck. This allowed me to separate the inner three decks a little so I could see better. This confirmed what I saw earlier today: The tab is melted away. I tried taking a photo of the damaged, but missed getting the focus right - and now I have it reassembled.

I did take a photo of the almost 50 years old ELCON capacitors. I have no reason to suspect they are not perfect, so that's great! Ripple looks perfect when the PSU is loaded so no sign of reduced capacitance or increased ESR, and they're keeping their charge perfectly even showing dielectric absorbtion, not self-discharging after use. One has apparantly run a little hotter than the other or the plastic they used to cover it was different or lower quality.

[adinsen]

I realized yesterday I've wasted a lot of time looking for a problem that wasn't: The previous owner told me the unit was not able to supply the rated output currents. I did some quick measurements that seemed to confirm it, but never checked it properly before starting to look for the cause of the issue. With no problems emerging from my checking, however, I decided to do proper measurements according to the procedure described in the manual. I don't have a power rheostat so my measurements aren't complete yet, but the output currents I measured are within specifications:

6V range:
I_Limit = 4.3A @ 6V output (1.2Ω load)

30V range:
I_Limit = 1.7A @ 7V output (4Ω load)
I_Limit = 1.4A @ 2V output (1.2Ω load)

I did the output voltage calibration procedure. I could calibrate the output voltages, but the relationship to the potentiometer setting is not linear. There are a few carbon mass resistors in the circuit and I should probably check them.

I'm still interested in hearing from people who have repair experience with decked switches as I have the melted one and the switch blades and rotors are all very dirty. Is it worth disassembling completely? I'm not particularly concerned with the 5 discharge resistor not being connected as the design around it is fundamentally flawed: Switching rapidly from 60V-range to OFF means discharging about 80V over 5 ~ 16 A. The switch will only survive that a few times. Also, I don't see its purpose and the failure of it is certainly not related to the over-voltage-protection failure.

I found a 2N3055 in a drawer so the over-voltage-protection is now working 

[Neomys Sapiens]

Further disassembly and attempts at repair of a decked rotary switch are only beneficial to the device being repaired, if
1. the rotary switch is from a series for which switches in kit form were (or are) available
or
2. if you are able to cannibalise another switch from the same series in order to gain the damaged part.
Else, you would be better off with outright replacing the switch. If you have identified a condition prone to cause such damage, you could attempt to migitate it by either a circuit modification (for example, add another resistor on the other side of the switch or increase the existing one) or by selecting a more robust switch for the replacement. Also, you could simply remove this one circuit from the switch, rewire it to switch the control voltage for a relay and use a suitably rated relay to do the actual discharging. One has to keep also in mind, that rotary switches in open construction are very susceptible to contamination and dirt. Maybe it was a minute speck of dust or, beware, a tiny flake of metal from another place in the unit which started the burning, which then progressed under the kind of load described.

[adinsen]

Thanks a lot for your advice!

I have found what appears to be a suitable rotary switch on eBay https://www.ebay.de/itm/184794218742 But with the power supply now working properly, I will of course not be replacing it. And the decks will probably not be compatible. Also, this is another open frame design with the same potential issues as you point out. I agree that it could have been a tiny flake of metal causing the burning. The PSU has open ventilation holes in the top, so the metal could also have come from the outside.

I agree about your idea of performing the discharge via a suitably rated relay (e.g. a 16A auto relay), but I am inclined not to implement it as I don't see the point in discharging the filter caps at all - I could be missing something, though!

[Neomys Sapiens]

Well, if you switch from Off into the 6V range, having been previously operating (worst case) in the 60V range, while demanding immediately a high output current, it is conceivable that the PS could run into a SOA or thermal problem for the output transistors. I think that the discharge circuit was intended to guard against such a problem, even if it occured only as a violation of Oltronix' own robustness guideline.

[adinsen]

You may have a point there! 

However, doing such a rapid changeover will also be likely to blow the over-voltage protection transistor. This is a 2N3055. The OVP essentially short-circuits the output of the PSU until the output voltage becomes less than about 6.7V (on mine). The OVP is connected via the swtich only in the 6V-range setting.

2N3055's has an absolute maximum rating of 60V V_CE. In the 60V range, the output of the PSU can easily be regulated past 70V. The problem is that there's a 4700 uF capacitor in the output, so changing back to the 6V range from a 70V output will connect the 2N3055 CE directly to the 4700uF capacitor in the output of the PSU! 

So that rapid-changeover is never safe, especially not when the output is unloaded... 

(I thought about this thanks to a comment on my photo here https://flic.kr/p/2ppcBei )

[adinsen]

A little update on my vintage Oltronix. I realized the A-meter was not showing correct values in normal mode nor with the 0.5A button depressed. I don't remember whether it worked when I did my first tests of the PSU or I just forgot to check that. But that deserved due attention, of course.

So a few days ago I started following the calibration procedure described in the manual. Unfortunately I couldn't get it within specs.

The A-meter is driven by a separate differential amplifier circuit measuring the voltage drop across the series resistors. It's a three tansistor circuit and while Oltronix has made good use of 2% metal film resistors in the overall design for critical resistances, this part of the circuit is made with standard 10% carbon comp resistors of the time.

I couldn't find any problems with any of the BC109 transistors, so I started checking them and they consistently measured to be between +7% and +15% of their nominal values. This may not seem too much off given their factory 10% rated tolerance but the consistent drift towards higher values was still enough to raise my eyebrows.

So I repleaced them with current, cheap but modern 1% metal film resistors (one of the great component improvements over the 50 years gone by!). That solved the problem: I was able to calibrate the A-meter within specs.

I haven't found a good solution on the discharge-connection problem, so for now I'm taking care to discharge the filter and output capacitors with a power resistor on the output before changing ranges, if I really need to.

I'm still fascinated that Oltronix designed this power supply to be a precision instrument when I was just a baby, and apart from the minor issues I've given attention to at the little TLC I'm giving it, it still is!