Armstrong 626 troubleshooting problem

audiogymp

Active Member
Hello AK

I'm trying to bring this British piece back to life but I'm a little stumped with the cause of the problem - it's a bit of an uncommon receiver. There was one thread on AK that suggested there might be a few of you that know the piece. I'd appreciate a little guidance if any of you have experience with it, or the inclination to help.

The schematic is available at http://www.audiomisc.co.uk/Armstrong/600/600faq.html. I've also included some photos.

There's no audio output. I've tested/observed the following things.
  1. Power-up on DBT - looks fairly ordinary: 28W bulb glows while caps fill and then dims to almost imperceptible glow. (Measurements were made using 70W globe, which only gives a dim pulse during power-up.)
  2. Voltage supplies for power stage appear to be just under 75% or expected voltages (57V at +ve side of main, 3300 microF filter cap instead of 82; 28V at +ve side of 4000 microF speaker caps, instead of 40V).
  3. Power transistors appear to be OK - OC measured E-C.
  4. There is some DC offset on outputs. The outputs are cap-coupled. I measured them with a 57Ohm load. They start off quite high as the caps charge but drop off quite quickly. Once they settled down there was a fluctuating voltage, generally up to 30mV (sometimes up to 50) across the load.
  5. Quiescent current was measured on resistor 226 (L and R). These are 0.68, not 0.47Ohm as shown on schematic. During power-up the voltage rises to about 13mV (just under the 13.6 expected for 20mA stated in schematic) but soon drops to zero as caps charge. I understand that the corresponding resistors (227 L and R) for the other outputs should read the same (albeit on the C rather than E side of the output) but they show no voltage at all on power-up or after. It is hard to see how any real DC voltage is coming from the output transistors.
  6. Resistor 23 (just under the rectifier in the schematic; 3rd photo) runs hot. It gets too hot to touch quickly, gives off a noticeable smell, and has darkened the insulation on the wires leading to it. I suspect that this is linked to the problem.
  7. The two caps that are parallel to a pair of the rectifier diodes are darkened (4th photo). However, DC voltage across the bridge takes expected steps of 0, 24 and 58V (although it should go up to 82V instead of 58V).
  8. The tuner section appears generally not to be working. However, on power-off (while it is running on charge from caps) it appears to start working - at least the LED indicator for locking into a station appears to come to life and respond to tuning. This is brief and only last while there is charge in the caps. This may also point to the cause of the problem.
  9. During the course of making measurements I tripped the earth return on the house fuse box a couple of times. The overload on the power board I used didn't respond.
Thew big caps have some bulging at the terminal ends and will need to be replaced. However, they measure OK on component tester and show OC on a multimeter. They don't seem to be the cause of the problem.

The function of the tuner on power-down and the transient quiescent current are interesting but I don't know what to make of it.

There is a cap on the mains side of the transformer (across the power switch). The component tester recognised it as a functioning cap (measured in situ) but the values were out. It doesn't look like the source of a short.

Any thoughts on where to go from here would be much appreciated.

Cheers, Jon
 

Attachments

  • IMG_0358.JPG
    IMG_0358.JPG
    164.7 KB · Views: 44
  • IMG_0359.JPG
    IMG_0359.JPG
    158.3 KB · Views: 46
  • IMG_0360.JPG
    IMG_0360.JPG
    114 KB · Views: 44
  • IMG_0361.jpg
    IMG_0361.jpg
    73.8 KB · Views: 40
Jon, did you make sure the power supply thermal delay is working fine? Looks like permanently engaged. Is that resistor heating up the one corresponding to the thermal delay circuit?
 
Hi Pablo (if I can call you that)

Thanks for such a quick reply.

You raise an interesting point. I think the resistor that overheats is the 47 Ohm one in the dashed box labelled "thermal delay". The schematic shows a component in parallel with the resistor that looks like some kind of switch - presumably something that is responsible for cutting power. If this "switch" were constantly open, one side of the AC feed to the rectifier would be forced to come through the resistor, which would lower DC voltage from the rectifier and overheat the resistor - nice thinking.

I have never heard of an arrangement called a "thermal delay" and wasn't able to see anything in parallel with the overheating resistor. I will look again. Are you able to give me an idea of what I might be looking for?

That said, I think there were 11.5V lines going to the attachment block the resistor is attached to - you can see it in the photo. I assumed it was just an isolated mounting point but perhaps there is something inside it.

Thanks, Jon
 
Hi Pablo (if I can call you that)

Yes. that's my name.

The thermal delay switch is correctly assumed the block under the resistor. When you turn the thing ON, this switch is open and the PS storage capacitor get charged through the 47 ohm resistor (to protect the rectifier bridge from over current). After a while the switch should close and full AC voltage should feed the PS. Clearly, the 47 ohm resistor should not heat up if this switch closes.

The solution is getting rid of the thermal delay circuit (bypass it), and replace the rectifier diodes with 1N5403 diodes; these will not blow. You may also install a 5A rectifier bridge.

After the mod you should get normal PS values, as stated in the diagram (82V, and 45V). Hopefully, you'll also get normal audio output.
 
Thanks, Pablo

I tested the theory by using an alligator clip to create a short across the 47 Ohm resistor several seconds after power up.

The voltages rose to 72V and 35V respectively. There is obviously a problem with the thermal delay but taking it out of the circuit doesn't restore full voltage.

Voltage and quiescent current at resistors 226 and 227 were still zero at idle. The tuner did not return to usual function but only appeared to work for the short time after power-off while there was charge in the caps.

I then tried a signal input with speakers connected. As I adjusted the volume dial there was a loud scratching noise and the bulb on the DBT went bright with the loudness of the noise. The light/noise was pretty much random while I worked the dial and no sign of input music (this was obviously brief).

There is something else going on here. Any thoughts?

Jon
 
Jon, may be there's something else going on, and may be not. You strongly need to have the power supply in perfect working order (with or without the thermal switch) and then you can move onto any remaining issue.

First thing, bypass the thermal switch permanently (until you find a replacement). Be warned, exchanging the rectifier diodes is advisable. Stock ones are too weak for the load.

Then, remove the DBT since, it's of no further value having it connected; plug the mains connector into the wall socket.

Turn the device on and read the PS voltages: transformer output (60 and 11.5 VAC), then both DC values.

Disconnect the speakers and connect a headphone set; this would make testing audio output more friendly. I've read in the link you posted above that volume pots in some Armstrong units do fail. Would be good practice to apply some specific Deoxit or similar lubricant in it to clean the pot's wiper and track. Also, do this procedure with all switches there. You don't want to search for an apparent issue that in fact could just be a dirty pot or switch.

Let me know about those PS values and we'll proceed accordingly.
 
Hi Pablo

Thanks for helping work through this. Here are my new observations.
  1. The thermal delay appears to be working. It might have been my attempt to remove/inspect or it may be that the reduced input with DBT was causing it not to function. I will keep an eye on it but no need to bypass/mod at present.
  2. With full supply, the DC voltages are 80.5 and 39 - so pretty close to spec.
  3. The AC voltages from the secondary windings are 62.2 and 11.7V - again close to spec.
  4. The quiescent current measurements across the resistors were 1 and 4mV on the left and 17.5 and 20.8 on the right (should be 13.6mV). Quite a spread but an interesting 3mV difference between the 226 and 227 resistors on each side. (On another thread I read that the voltages across the 226 and 227 resistors should be the same).
  5. With music into the aux input and listening through the headphones there is no sound - just low level hum when the volume is high. Interestingly, after power-off, there is a short period of faint music as the caps discharge - easily noticeable on the left but might also be faintly there on the right.
What is described above is the state of the amp after I was finished. But something interesting happened in the course of making measurements.

I initially measured for the 60V transformer output at the 47 Ohm resistor we discussed (between it and the chassis/ground). This measured at 75V. I'm pretty sure that I was measuring AC because I was surprised it was this high - perhaps I had the meter set wrong but I double checked. Then I made some other measurements and the ground return on the house fusebox tripped. After I reset it, I measured again and the AC at the 47 Ohm R was at 36V. I then realised I was measuring the voltage across half the bridge and found the other secondary line from the transformer and measured it correctly.

Perhaps the 75V reading was an error but this is the third time the earth return on the house fusebox has tripped while I've worked on the amp. Most of the other voltages above were a little higher before I tripped the earth return - but that could be because the amp was warmer when I measured them before.

Cheers, Jon
 
Jon, your incorrect AC reading was due to the fact that the transformer output is not connected to ground. You need to read it on the transformer wires or at the right spots on the diode bridge.

Good news that DC voltages are normal. Now you need to know what stage is not working. Very rarely both channels fail so it's indeed weird what's going on.

Quiscent current ideally should be the same on resistors 226 and 227, unless some parameters here and there have drifted causing the difference you read. May be an issue hiding behind or not. We'll see later.

I'd like you to inject audio signal at S27 (loudness switch) on the terminal leading to the volume pot, or directly on the volume pot. Use the volume pot ground terminal for input signal ground. Audio output should vary with volume pot adjustment.

Please do this test on both channels. If you get audio output we'll troubleshoot the pre.
 
Hi Pablo

I did the test you described. Both channels produce headphone output and the volume can be controlled with the volume pot. I take it that this means that the pre needs to be checked.

What would you like me to do next?

Thanks, Jon
 
Cool! The amplifier section works. Last step will be adjusting bias level back to spec.

Regarding the pre, please inject audio signal in either terminal of S22 L.F.Filter. Attach signal ground wherever to the chassis.
 
Hi Pablo

I get sound through the left channel but not the right channel. The good channel sounds OK.

I might get some multi-turn variable Rs before I set the bias. I read somewhere that when these Rs go OC, the quiescent current goes to max. Do you agree? and should I start with them set to min resistance when setting?

Thanks, Jon
 
Interesting. Jon, what voltage value do you read on the diodes' cathode side of 521/522/523/524? Expected voltage is 12V.
 
They are all around 12.6-12.7V - close to 12V stated on schematic.

It's a little difficult to navigate the circuit boards because there are no component numbers. Also, some of the components don't have the stated values (e.g. the 2.2 microF caps near the diodes were 0.22). But would you like me to check voltages against the schematic (as far as I can) and report back?

Jon
 
Jon, those caps you mention need to be 2.2uF (1 to 10uF would be fine). At 0.22 almost no input signal would go through them. Is there a chance someone replaced the originals with the wrong value ones?

Please try injecting audio signal at any of those diodes' cathode. Since you have 12V there inject signal isolated via a 1 to 10uF capacitor (capacitor's positive terminal to diode's cathode). Signal ground to chassis.

No need to check voltages yet.
 
Pablo, I'm sorry but I made a mistake with the previous measurements. I was measuring at the anode side, which we would also expect to measure about 12V.

I re-measured at the cathode side and the answer is different. At first two of the diodes had a voltage at the cathode side (left AM and Tape, I think) of about 13V. I then remeasured at the anode side - unlike before, all measured zero. Then a re-measured at the cathode side and all measured zero. At one point I was able to hear a faint sound on the left side from the signal injected at one of the input sides of the diodes (I think the AM).

Then something interesting happened. I had left the receiver turned on while I left the bench (I was away from the unit) and the house earth return tripped. I turned the amp off (but left it connected to the mains) and reset the fuse box. The fuse box tripped again after about 20 seconds. I switched off the power board the amp is plugged into and reset the fuse box - it was OK after that. I then turned the amp on again and checked DC supply voltages - it appeared to be working as before.

There seems to be some intermittent leakage to earth going on. However, the main supply seems to be operating as expected.
 
Strange about the earth leakage, Jon. You may install you DMM set on mA between the earth connection in the wall outlet and the amplifier. Sounds like you need to measure above or close to 30mA to make the differential switch trip.

Regarding the 45V power supply, do you have a consistent value there?

Check out in the diagram, the input selector is fed with 45V from that PS; then through a couple resistors voltage is reduced to 12V and applied to the selected input diode forcing it conduct. The input signal travels through the diode sort of modulating the 12VDC value. Signal is recovered at cap 494.
Since any given diode is receiving 12V, which one depending on the selected input, and since all cathodes share the same path, you should always read 12V at any of them (cathode side), being on the selected input or not. Switching between inputs should not alter this. If you don't read 12V there then the 45 PS must be faulty or the input selector damaged.

Please let me know your findings.

Amp2.jpg
 
Hi Pablo

Thank you so much for explaining the signal input arrangement and taking the time to mark-up the schematic. I've learnt a lot from you instructions, which helps me take a little initiative with the work.

My observations:
  1. voltage at the source selector at supply input is 47.8 (cf 45V) - about right.
  2. The source of the supply is transistor no. 628. The voltages are C=78V (cf 75V); E+ 47.8 (cf 45, connected to input selector); B= 48.4V (cf 45.5V).
  3. The voltages at the diodes in the input circuit are about right when Tape, Aux and Tuner are selected: 12.7-13V at cathodes of all; 13+0.5 at diode pair of selected input.
  4. When Disc is selected, this changes: 5-5.5V at the Cathode of all diodes; 5.5-6V at the Anode of the Disc diode pair.
I now understand why was getting different measurements at different times.

The earth return tripping happened again just as I described before - including the second trip with the unit turned off but still plugged to mains. But then I was able to work on it for a while without a problem. It isn't easy for me to test earth current but the use of earthing in the circuit makes me think the earth return issue could be related to the problem we are looking for.

After typing this, I thought that perhaps some of the inputs should work and I may have missed this before. I tested inputs again, listening with earphones, but there was nothing but a slight, distorted signal on the left side coming through on them.

Thanks again, Jon
 
Jon, please inject audio signal on the spot marked by the arrow below. Stick to the working channel; we'll take care of the other soon.

Amp3.jpg



Also, regarding disk input, select disk and check out the voltages marked. An unexpected voltage drop of any of them suggests a bad contact (input selector switch), a damaged resistor, or a bad solder point. If the switch is fine measure directly on the resistors. If they read fine there is a bad solder below.

Replace the capacitors maked green with the correct value ones (2.2uF/16V)[Alternatively you can install caps ranging from 1 to 10uF/16V or 25V]



Amp4.jpg
 
Hi Pedro

Thanks for the further instructions.

It's not an easy amp to work on without circuit diagrams because the PCBs are not well labelled and some components differ to those in the schematic. It is also compact with short wires and poor access. I will need to cut and extend wires so I can see the conductor layouts. I will get onto the measurements again tomorrow.

In the meantime, here is a photo of the circuit board we have been looking at. The black/yellow/rectangular components from GEC are the ones that I assume are 0.22microF caps in place of where the schematic shows 2.2mF. They all appear to be the same and I don't see 8 of any other capacitor in this part of the amp.

Would you please confirm that I am correct that these are in fact 0.22mF. If so, would it be possible to leave replacement of these to the end and possibly reconsider then? I am almost certain they are the original components and the unit must have worked with these caps before it failed.

Cheers, Jon
 

Attachments

  • IMG_0362.JPG
    IMG_0362.JPG
    118.6 KB · Views: 28
Those square ones are as you said 0.22uF. Only after removing the board you'll be able to tell if those are the input caps. There are some other 2.2uF electrolytic ones there so who knows right now.

Let me know what you find out regarding those voltage values marked above.
 
Back
Top Bottom