Noisy Hackers!

[Mark Hennessy]

Hi All,

As you know, I pretend to know a thing or two about amplifiers. It's not true really, and so I thought I should share this with you - just so you know that even so-called experts sometimes get caught out

Many Hacker sets are noisy at minimum/low volume. The noise is quite distinctive - rather than the hissing you might expect, this is more of a rumble. That's the first clue - this is "shot noise", or 1/f noise, rather than the white noise that thermal noise causes. Thermal noise is normally built into the design, others are to some small extent, but normally indicate a faulty component.

In my experience, the "Lockfit" transistors that were widely used in the 1970s are relatively unreliable, and are prime candidates for causing this problem. Next up, the DC offset preset can very occasionally cause the same symptom, and it should be cleaned and wiggled first, because that's cheapest . Finally, if you still have noise after changing the first couple of transistors, you have to think about changing carbon composition resistors...

Anyway, the set in question was a Hunter RP38A - although Hacker used very very similar designs in most of their sets. I've attached the schematic.

So, wiggle RV3 - this helps a bit. Good. Change T1 for a BC108. Yes, I know the schematic says BC108, but invariably they come with BC148. Guess the Lockfits were cheaper than BC108s, but BC108s are definitely more reliable...

Normally, this would fix it, but of course, it makes no difference. No matter - sometimes T2 can play up. Not in this case :-(

Consider T3. After all, it's a dreaded Lockfit. But as you move forward through an amplifier, the later transistors have less opportunity to contribute. Still, the drawer full of transistors is still on the bench, so in goes a BCYsomething. No difference.

Damn! Looking at the pile of innocent Lockfits, I say "sorry to have suspected you all", and back they go. What could be causing this? Then a distant memory flickers, and I remember that R1 (150K) has caused problems in the past. Though not this time, obviously :-(

Time to panic. Consider changing the remaining transistors, but know that's crazy, so decide to park that idea. Consider doing a blanket swap of all the resistors, but that's similarly silly - yes, it's almost certainly one of them, but I wouldn't learn anything from that approach, and it's not exactly sympathetic. Oh well, in the absence of a better plan, decide it's time to play with test gear. That'll make me feel better

I have an ancient Lindos LA1, and while it's not much use for serious R&D, it's perfect for old radios; I use it a lot for this sort of thing. I'm able to determine that the noise from this set is more than 10dB worse than other sets I can quickly lay my hands on, so not allowed to give up yet!

In a negative feedback circuit, you have a chicken and egg situation, and as it's impossible to break the loop outside of Spice, it's nearly impossible to narrow down a faulty component by measurements. This problem was my excuse for spending ages blankly staring into middle distance for an hour or so...

This amplifier has two negative feedback loops. All stages have DC feedback, but the AC feedback only encompasses T2 onwards. If it didn't, the tone controls wouldn't work!

And, there was a strong clue - the LF "shot" noise was strongly affected by the bass control, suggesting it was indeed coming from the first stage. Putting the Lindos on the top if C3 and C8 proved the caps were OK, and not passing noise into the first stage. So, suspects are T1, R1, R3, R4, and possibly R5. That's more like it - I've already changed T1 and R1, and it's probably not R5. Change R3 and R4, and Bingo! The set is 5dB quieter than my known-good example! At last!

It can't be both of these, surely, so replace R3 and the noise is back. Put the replacement back in place, and everything is good again. Put the original R4 back in place, and no change. Fine, that one can stay. Originality and all that... Finally, for laughs, I revisit R1, but the original component does raise the noise floor a bit, 3 or 4dB. That stays out then...

This should have taken a lot less time, and I'm not sure how much I've really learnt, because most people know that carbon composition resistors do go noisy, but it's a random thing, and the chances of another R3 failure are quite slim. Still, it does one good to be caught out by a "simple" fault from time to time :-)

The resistor in question is one of those brown box ones, with two wires emerging from one end, and 4 strips painted on the other. Widely used in the early '70s - no idea what they're called...

Anyway, amplifier is all back together and set up, just need to do the RF alignment (as always for a Hacker, it seems). Another excuse to play with test gear, but that can wait until tomorrow.

Hope this is of interest, and apologies to Lockfits everywhere

Mark

[tiv48]

not just me then who has a mistrust of lockfit transisistors, not just in audio circuits but in colour decoders and tv if strips,
mainly noise but also varable gain . i used to give them a blast of freezer .but if suspect change it its the only way. no good testing them most will pass.
secondly what was wronge with transformer coupling for small amps ,no mass distruction of transistors
rob t

[Mark R]

Interesting write up Mark, very informative.

cheers Mark

[Mark Hennessy]

Hi Rob,

Yes, the BF194/5s are used extensively in the RF and IF sections, and can indeed fail. So far, they seem slightly more reliable than the BC148s, but I guess there are factors at play - first, the sort of things the transistors are exposed to (collector currents, operating temperature, environmental stuff), and next, the impact the transistors are able to have. A slightly noisy transistor at the input stage is ruinous, whereas if that same noisy transistor is being used as a driver, or perhaps a VBE multiplier, then that noise wouldn't be noticed. Indeed, when trying to preserve originality, I've been known to swap transistors around rather than replace from stock.

The freezer tip is a good one - heat from a soldering iron can also isolate them. Incidentally, freezing that faulty resistor make no difference to the noise level...

As for transformers in audio amps, I accept they make things a lot simpler to understand . But, they are quite bulky and expensive - well the good ones, at least, and they do add distortion - allied to that, because of phase shifts, with an amplifier with transformers, you are limited in the amount of corrective negative feedback you can add. Given that designers are tasked with getting good performance at low cost, you can see how these things evolved...

All the best,

Mark

[Joe]

Time for me to show my ignorance again: What's a 'Lockfit' transistor?

I found a hair-drier blowing through a funnel a good way of finding dry joints on my oscilloscope - first time I've used one for years.

- Joe

[Mark Hennessy]

Time for me to show my ignorance again: What's a 'Lockfit' transistor?

It's a type of transistor package - have a look at the picture; you should see three of them.

Rather than the normal 3 wires emerging from the bottom, there are 3 pins with protrusions that are designed to keep the transistor firmly in place during soldering...

All the best,

Mark

[Joe]

Thanks for the explanation. I have seen them in things.

- Joe

[pwdrive]

We used to call them something else...reverse the lock & fit and sub another word for fit.

The subject of transistor problems has taken me back to my days with Sony in Canada.
Back in the seventies at Sony we were starting to get stuff in with various faults but a lot of them seemed to have a familiar "ring" to them, most commonly intermittant problems (the worst) across all types eg: amps, tape decks, tv's etc, some of these were real nasties ie: iintermittant only once in the w/shop...send back out...bounced a few weeks later.
Typically (but not always) the symptoms would be intermittant crackling or noise (like resistor noise) in audio products or intermittant video/IF/chroma problems (colour tv's) it was giving us a real headache.
First approach was the normal one, intermittant = dry joints, out come the blunt end of screwdriver and a knitting needle, bang bang tap tap till your in the area, inevitablely it was the print side of a transistor connection, got to be a dry joint then, out comes the bench magnifier...looks ok but I will resolder the three leads anyway...stick on soak (sometimes for a fortnight) it's fine, send back, next week it bounces...sod it, same proceedure again only find it on a different transistor, same fix, same soak same bounce.
Ok time to nail this little b*****d had a real good luck and a re think as we were now getting a lot of stuff coming in with weired faults and the boss is getting concerned about the back log, normally I'm on the fault within half an hour at most, this was getting to me, anyway as said had a re think, maybe the transistors are responding to taping due to a problem with the transistors internall connections, takes transistor out and checks it, all ok. Eventually a tape deck came in that was intermittant noisy on audio and the intermittant problem stayed long enough for me to find what was causing this problem with the transistors used in most of the Sony products, it responded to the usual taping etc, again narrowed it down to the area of a transistor, wiggles the top of the transistor and the fault aggrevates, must be a dry joint or an internal failure, it was neither of those things although the transistor was to blame, I removed it and had a real close look under the bench magnifier at the base of the transistor (the underside) and there radiating out and under a certain light was flourescence looking deposit (looked like what petrol on water would look like) it radiated out from the center lead (I think that was the collector) and had spread to the other leads and it was conductive, so taping the PC board was enough to cause this conductive connection to become none conductive due to the vibration caused by tapping or wiggling the top of the transistor, the act of removing the suspect with a soldering iron also broke that conductive link but given time, usually a couple of weeks, it would be back to its old tricks again. From that day on finding these nasty intermittant faults caused by the above became a breeze, just slighty stroke or wiggle the top of the transistors and it soon became clear which one was at fault.
When I eventually returned to the UK I needed work so I got a job as a bench tech with a local Radio & TV firm, most of the techs were only used to the usual Thorn, Phillips,ITT stuff, they asked what my favourite stuff was and I told them Sony or other Jap stuff..."Right then you can repair the bl***y lot" which was fine by me as I knew the problem that had probably been giving them grief for years, once I showed them the trick they were queing up to fix them.
The above transistor problem was also found on Hitachi transistors as well as Sony ones, the fault was also present on the semi round type transistor encapsulations as well as the that had rectangular encapsulations, that led me to believe that the conductive deposit some how appeared from something to do with the PC board proccess or it was coming from within the transistor itself or some kind of coating on its leads, it was always on the center lead and radiating outwards, it was dry, scraping the base round the center lead would also cure it if desperate but transistor replacement was the only sure fire cure, but at least I had found a quick way to diagnose the problem...wiggle or stroke...
All that was back in the 70's.

Lawrence.

[Joe]

Thanks for that post, Lawence. We could do with some more Tales from the Trade in The Lounge.

- Joe

[pwdrive]

Thanks Joe....Yeah, the good the bad and the ugly, I could tell the one of how me and one of Sonys top engineers from japan solved an early Betamax Video Cassette Recorder production problem, it beggars belief but it's true, I will post it up as soon as I can, certainly within the next few days if not sooner it will be a bit of a tease, I will start the post with the symptoms and see if other members have any ideas on the said symptoms...

Lawrence.