Germanium transistors Farfisa Duo Compact/Ballata

[tardishead]

I just bought a Farfisa Ballata for 1Euro
It's fantastic and is very close to the sought after Compact Duo which is the early Pink Floyd sound amongst others.
I have a manual but some of the circuit boards are different and predate the manual.
The lower manual keyboard sounds are a lot quieter than the upper manual and bass pedals. In looking into the mixer/preamp circuit board I discovered that the V+ is very low. 5.5vdc
Most circuits in Farfisas of this era run on 8v. Never have I seen as low as 5v .In stead of taking the V+ from the 8v rail which feeds the signal generators and other circuits  - the Ballata takes the preamp/mixer v+ from the 20vdc which feeds the reverb driver and speaker amplifier by adding a 820R resistor and a cap. After this resistor I measure 5v which implies 18ma current draw. If it was a 10v drop that would mean about  12ma.
The preamp mixer circuits consist of five oc76 germanium transistors. I checked all the resistors and they are pretty close to tolerance so it would seem to point that the transistors are drawing more current than intended.
My question is could increased leakage over time account for this larger current draw? What is the best thing to do?
Simply lower the dropper resistor for less voltage drop?
Or replace those transistors?
It seems to sound good BTW.
But obviously raising v+ would improve headroom and noise issues. Is raising v+ likely to kill those transistors?

 

[PRR]

Most of the transistors must have collector or emitter resistors. Look for one showing much more voltage drop than the rest-- that transistor may be leaky.

Also always suspect rail caps. The go from 0.1mA leakage to 1mA leak to 3mA leak, which seems to be about how much you are short.

 

[tardishead]

Yeh i change all the rail caps always
Can transistors alone collectively develop so much leakage?

 

[Bo Deadly]

Yes, OC76 is an old(er) "grown" type of Germanium which are known to have bad leakage. Howewver, I have not heard of transistor leakage increasing over time. If a Ge transistor has high leakage, it is more likely that it was always leaky.

So you might check other things like PRR suggested replacing caps.

If that doesn't fix it, just debug the circuit like normal. Trace the circuit out carefully, study it and then look carefully at the circuit and see if it's working right:

Is the bias right? Put a larger signal in and see if it's fully utilizing the voltage swing or is it being asymmetrically clipped?

Is the amplification factor correct?

If there is a problem with transistors it's probably one or two and not all five. Either it's just really low gain or high leakage. As PRR said, you can measure the voltage drop across applicable resistors (the emitter resistor as long as there isn't significant contribution of current from other resistors also connected to that emitter).

If you do feel like you want to try to start changing transistors, bear in mind that Germanium transistors are very sensitive to be being heated. If you put a soldering iron on a pin for longer than ~5 seconds, it will completely destroy the transistor. And you should save the transistors because a really good "grown" germanium transistor could be worth $$$ (to use as the first transistor in a Fuzz pedal). I don't know what the amp card looks like but if it looks like you can remove the Germs without frying them you might elect to just remove all of them and check them for leakage and gain.

Leakage affects the circuit more in bias sensitive locations. The first transistor in a Fuzz pedal for example must be low-leakage (<100uA) but the second transistor doesn't matter as much.

So my point is you might be able to just re-arrange things to get it to bias properly.

If you find transistors with really low gain or you don't have suitable low-leakage transistors you can by NOS on Ebay. However, virtually all of the "grown" Germanium transistors on Ebay are going to be horribly high leakage. And you cannot believe claims of low-leakage. I have purchased many Germs from Ebay that claimed low leakage and they were not. However, "alloy junction" transistors can be low leakage (maybe too low depending on the circuit) and I think they would be fine for your keyboard amp (although I am not familiar with that keyboard and certainly not the specific circuit whatever it is).

This Ebay seller:

https://www.ebay.com/sch/m.html?_odkw=pnp&_ssn=orpheus_2005&_nkw=germanium+pnp

from Bulgaria (used to be called bgebus) sometimes has Icbo measurements of the actual transistors for sale that I have found to be close. To calculate leakage multiply Icbo * HFE. For example 160 * 2.2 = 352uA which is too high. But 60 * 1.4 = 84uA is good.

All of the "SFT" parts in the link above will be alloy junction (you can tell by looking at the case). Note you can lookup transistor characteristics at alltransistors.com including old Germaniums.

So if you have a low-leakage OC76 with enough gain (>60), use it for the high gain transistor and replace any "bad" transistors with something like the aforementioned SFT or similar. Unfortunately it is likely that you will not have / find an OC76 with low-leakage / high gain in which case you should put one of the SFT transistors in the high-gain position.

To remove transistors I would attach an allegator clip to the lead being de-soldered, put the iron on high, add a small amount of 60/40 Tin/Lead solder whist heating for <2s and then suck it out with a giant solder sucker. Then move the heat-sink clip and repeat.

Or it could be something else entirely like a weak power supply or bad caps in which case forget everything I just said!

 

[tardishead]

Thanks for that amazing info
In the end it was a AC128 in the reverb driver circuit that had gone AWOL and was bringing everything down
Its a strange scheme for the mixer/preamp V+ but it seems to work fine.

 

[Bo Deadly]

Ahh, ok. Glad you figured it out and that the solution wasn't too painful.

AC128 is an alloy junction low-gain high(er) power type. Power Germaniums are particularly worthless as they usually have high leakage, they're sensitive to temperature and can easily get too hot. If they get too hot it will increase the gain enough that it will go into "thermal runaway" and basically self destruct. Fortunately we came up with Silicon transistors which are far less susceptible to these problems. If you really want to replace that AC128 try to find a bunch that are not picked over (v. difficult) and pick the one with the lowest leakage (low leakage for a power germ could be well over 100uA) and add a heat sink somehow (some AC128 have integrated heat sinks with a hole for a bolt).