Cannot bias SAE A202

[genlabs]

Hi all:

Full disclosure: I am no expert at this stuff. I have repaired my fair share of equipment but most of those repairs were straight forward, easy to find parts repairs. I am trying to learn at least one thing everyday.

I am working on a SAE A202. I cannot bias this thing at all. The device plays but at low volume there are some artifacts. The heat sinks remain cold. I started Googling information about biasing diodes and blhagstrom pointed me in the directions of some mind blowing information on how members of this forum were constructing bias diodes from common diodes in a string. I thought "cool this is where I should start".

The bias diodes in this amp are super fragile so I did not want to remove them unless absolutely necessary.
If I measure across them in circuit, with my Fluke in diode mode, I read @ 1.63 volts.
If I measure voltage drop across them while the amp is powered up I read @ 1.9 volts.
I was really ramped up about building my own "homegrown" bias diode but after taking these measurements I think my bias diodes may be ok.

If I put my DMM across the emitter resistors of the two adjacent output transistors I read absolute zero. Turning the pot makes no difference. I have no documentation for this amp. I have done rudimentary checks across the few components on each amp module. The resistors have drifted a little but nothing dramatic. Any ideas or clues?

 

[cademan]

If it's playing low in both channels, I'd probably start by looking in the power supply area.

 

[elnaldo]

Can you post a schematic?

As suggested, check power supply voltages first, check all the + and - rails are OK.

Then, if PS is OK, you need to measure at the driver transistors, and the bias transistors or diodes (those biasing the driver's bases) the same thing you did with the output emitter resistors (measure voltage drop, to see if the transistors are conducting). So try to measure voltage drop at the emitter resistors on the drivers and bias transistors. We want to know if they are conducting.

But check supply voltages first.

 

[genlabs]

If it's playing low in both channels, I'd probably start by looking in the power supply area.

Let me clarify a little; It is not playing "low" in both channels. I can make the amp play normally. Playing at higher volume it sounds ok. When quiet passages are playing I can hear some artifacts that are disturbing. When I say quiet, I am talking so low you have to keep your ear near the monitor. Maybe most would not hear it. I certainly can. Playing this amp against my reference ML amp, with the same source material, I can tell the difference. It probably not fair to compare the reference amp and the 202 but I am very sure SAE would not have put out an amp that had this low signal flaw. The best way I can describe it.......it sounds like the waveforms are being quantized or squared off. Almost like clipping but in the opposite direction. That is what led me to believe the amp was all out of alignment. I can dial in DC offset and it is rock steady at zero. Bias however has decided it was going to take a vacation in the Bahamas.

As for the power supply: I keep this thing on a Variac dialed back to about 70 volts while I am probing around just in case I do something stupid. I have been known to do that on a regular basis.
With it dialed back it had just under 50 +/- rails.
If I plug it directly into the wall I get 63 +/- rails.
These are values that I think would be typical.
The bias remains broken whether I am on the variac or not.

Thanks

 

[Powertech]

If you are hearing this 'noise' at very low levels and you appear to have zero bias current on the output transistors, what you might be hearing is crossover switching where the output transistors are working in true class B mode and are actually switching off at waveform crossover instead of just remaining on as they should be. This is normally termed 'crossover distortion' and it is what the transistor biasing is designed to reduce (usually to near zero). This is what makes the amp class AB instead of class B.

 

[genlabs]

Can you post a schematic? As suggested, check power supply voltages first, check all the + and - rails are OK.Then, if PS is OK, you need to measure at the driver transistors, and the bias transistors or diodes

Ok folks, we are going to go down the path of ironic humor. There is no other path. If you can't laugh at your own stupidity then why go on right?
Not just 5 minutes ago on my previous post above I proclaimed how I sometimes do stupid things.
Here is the short sweet story. The amp is on my bench facing the other direction. I can't see the indicators because I moved my mirror to let my wife use it.
I have the amp on a variac. I wanted to get true rails +/- voltages to respond to cademan. I plug the amp into a regular outlet.
I got the measurements, walked to my pc and posted my findings.
Decided I wanted to take a photo of the amp module board layout for elnaldo, since I did not have schematics etc for this amp.
Looked at the variac's lighted switch. It was off....good ! Moved the amp module and sparks begin to fly as the emitter resistor of one channel touches the heat sink of a driver on the other channel.
Why did this happen? I forgot I was not plugged into the variac anymore. Two output transistors are gone. No tears, just moving on. I am not ashamed to say how stupid I am.
I pulled the 4 output transistors and am ready to move on working with the remaining channel. In some ways this is good. I consider it the cost of my continuous education.
In trying to help myself I find ways to hurt myself.

Now I am worried about sourcing authentic Toshiba 2SD845 and 2SB755. I really don't want to test the ebay waters with their flood of counterfeit crap. Any reliable sources are appreciated.

Here is a picture of my wonderful arc strike on the heat sink.>>>

Channel

• genlabs
• Feb 24, 2018

SAE A202

Also @elnaldo this was the pic I wanted to send you to show the topology of the amp just before I went off on my arc adventure. The four outputs on the bottom and I believe two drivers above Sanken 958 and Sanken C2168. I believe anything to do with biasing is located on this board. There is a smaller input board hanging off to the side with a 5 wire umbilical attached to each channel. DC bias is done at the input board.

 

[genlabs]

If you are hearing this 'noise' at very low levels and you appear to have zero bias current on the output transistors, what you might be hearing is crossover switching

This is where my thought process was headed. I can't say I have ever actually heard cross-over distortion before. In my head I can imagine what it would sound like. I probably need to force future projects into non-normal working ranges to experience more of this on purpose.
I thought the ice cold heat sinks meant these output devices were barely on most of the time. I needed to push them way further towards A than they were but no dice when moving the pot.

 

[blhagstrom]

FWIW, my 202 is the dirtiest sounding amp I have ever owned.
On typical run of the mill decent vintage speakers, it sounds excellent playing rock.
I love it.

On very high efficiency speakers, (108db), the noise floor is astounding.
Now, I know things were still sketchy after I managed to get it working, but it sounds so good, I just run with it.

I checked my salvage stash, no 2SD845 or 2SB755, but I have 2SC2921 which is supposed to cross to the 2SD845. The crosses for 2SB755, I don't have but the 2SC2921 I have ran with 2SA1215, I have those. They are stronger than the crosses for 2SB755, higher current.

We already have a PM going so if these help, let me know.

 

[genlabs]

I checked my salvage stash, no 2SD845 or 2SB755

I appreciate you even looking. I think if am going to install cross devices I would want to do them all. I don't know if I am that committed to this amp. I have a 502 which over-shadows this little guy. I was thinking to get the 202 in top shape for a bi-amp project at some point. I am in no rush to get this going at all. Thanks for the offer!
I did see a few old stock used pulls on-line that may suit me.
All the newer Toshibas have the block lettering. The older stuff had that script font.
When I see that new style block lettering I can only focus on the fact there is a high probability that they are low quality fakes.
I don't know how rampant counterfeiting was in the 70's and early 80's but I would not be opposed to an "old-pull" from that era.
I guess even "old pulls" can be counterfeits in disguise. This is why I need a reputable source.

I concur with you on the SAE noise floor. The 502 is quite good, considering it was consumer level gear. SAE offered a great value when it was a viable company.
I look at he latest $20,000 SAE / ATI monsters kind of drooling, knowing I will never own one of those. They don't have that "magic", dirty something, that their value lines did either. Their old aesthetic is timeless. The new stuff...meh.

My 202 is limping on one leg right now but here I am listening to it as I type this. I will get her back on two legs again. And maybe even get those legs biased. I even like the way it smells like old electronics. I need to see a shrink.

 

[Powertech]

A word to the wise about Toshiba power transistors. I am reliably informed that they stopped making most of them around the year 2000 (2SC/2SD series). Most Tosh power trannies on the market are fakes.

 

[blhagstrom]

Let me know. I have lots of the ones I listed.
All pulls, all tested.
Since I pulled them, they should be originals.

 

[genlabs]

Most Tosh power trannies on the market are fakes.

Thank you.....I totally agree. This is why I would like to get a pull from some vintage piece. I think my chances are better with old grungy used stuff.

 

[Powertech]

You are probably right. Good call.

 

[genlabs]

So I am pushing forward with this amp and the bias problem. I can still perform exploratory tests with one channel out of commission.
I wanted to share what I found.
I don't have any schematics, so in the interim. I took photos of the front and back of the amp pcb.
I brought those into Coreldraw, flipped them around and made an overlay so I could easily trace out components.
Now with a road map I could work a little easier.

The rail +/- is sent into two amp pcb's and then it is also sent to a single input board via two 5 conductor umbilcals.
The 5 conductor umbilical is identical for the right and left channel amp.
2 wires are rail +/-
1 wire is negative feedback (labeled NFB on PCB)
I am not sure if I can define the last 2 of the 5

I focused on the wire that went directly to the bias pot. I was doing point to point measurements and the voltages were all over the place. So I let out a sigh and I noticed the voltage for the bias wire jumped high quickly.
Mind you I have the board in a jig, outside of the frame with the foil side facing me. I blow on the traces and again the voltages start flying around.
I don't imagine there is any temp compensation going at the input board. The temp sensing diodes are far away mounted to the heatsink.
I grab a drinking straw and start blowing on components. They are laid out kind of tight but I think it may be the 2SK333 FET Diff Amp.
I happen to have two of these so I will install them tomorrow.

Has anyone seen behavior like this? I mean there is a crazy swing from -5 to 30+ vdc if I blow on it for a second or two.
I can see this voltage swing appear on the base of the driver transistors.
I can understand some components changing value slightly if cooled but a swing like that seems nuts.

I am working on the channel that has no output transistors mounted. I did not want to risk taking out any more transistors

Any thoughts really appreciated.

 

[genlabs]

I swapped out both 2sk333 diff amps on the input board.
I still have only one channel connected to the input board. That channel has no output transistors mounted.

Now I have a steady voltage on the bias circuit but I think it is extremely high. It is sitting at rail voltage.
This is the voltage I am reading at the base of the driver transistor and the base of the output transistor pads (on the solder pad, no output transistor present)
I do not see any fluctuations in the voltage if I blow on it like before.

I need help. The more I play with this amp the more I realize I really don't understand what is going on.
What type of voltages should I see on the base lead of the drivers and outputs?
Are the voltages there so high because the outputs transistors have been removed.....lack of feedback from no output transistors????

I really don't want to solder in output transistors and see them evaporate when I power up because the base voltage s sitting at an astronomical value.
Any resident experts here who can chime in please?

(Photo edited to include warnings)

 

[elnaldo]

Hello. Please take it easy and relax, we can fix this small board if we are careful and proceed step by step. Worst scenario would be to replace ALL the components , what could be done in one afternoon, it's a small board.

Don't install the output transistors until the voltages are OK. Expect around 0.5V at the output transistors bases. It should react to the bias trimpot.

What I'd do:

To clean the pads where the output transistors should be. You need a solder wick to do that.

Then, clean with the wick every other soldering point you have made. You need to trust 100% there isn't any solder bridge shorting 2 traces, or any broken PCB trace. If needed use a magnifying glass.

Then, test the Driver transistors (you can do it in circuit), with the meter in diode mode, check they are not shorted. If in doubt, replace them both. A new transistor is cheaper than the time to test it. Check or confirm if they must be insulated from the heat-sink. Rail voltage at the bases indicates a short, or the driver, or other component and the alectricity find it's path to the driver's base. Check the emitter resistors are OK (resistors connected to driver's emitters)

check the bias diode. If it's a STV-3h or 4h, expect a couple of volt voltage drop, you can't measure it with a regular meter. You need a battery and a resistor in series with the diode, or just test it in circuit later, measuring the voltage drop across it.

Check the bias trimpot is OK, resistance across it, and resistance moving smoothly when you rotate the screw.

Connector labeled "A" is directly connected to your bias trimpot, check the components on the other board. To have bias current, you need a current flow through several transistors and resistors, you can confirm the current is flowing measuring voltages across the resistors. If some resistor is getting too hot, too much current there. The bias circuit handle low currents, so nothing should be too hot when the amp is idling.

Then, proceed to test the voltages, check for +V and -V at the collector tabs of the output transistors, aprox 0.5V at the bases, and near zero at the emitter tabs.

 

[genlabs]

we can fix this small board if we are careful and proceed step by step.

@elnaldo >>> Thank you for your input. I am sorry I did not respond sooner. I am working on a server project at my job and I am coming home like a zombie and falling asleep.
The scary looking pads were easy. I have a rework station. Cleaned them up in a few seconds and even tinned them a little. I really did not want to mess with the pads. I can tell these transistors were replaced before I received the amp. The previous solder job was kinda sloppy. I am always scared of traces lifting on older gear. It looks fine now.

Previously I tested the driver transistors in circuit. They tested fine. I will make it a point to check them again tonight just as a safety measure. The driver transistor heatsinks are wired "hot". No insulators. The mounting screw for the heatsink is in contact with the trace for the middle leg anyway (collector). The trace is wide enough that the screw head does not touch any adjacent traces. This is identical for both drivers.

I had a lot of curiosity about the bias diode. At first I thought it was thermistor. To understand the bias diode a little more I took four 1N5402 diodes and chained them together. I did see the voltage vary as I heated the diodes with hot air. Cool to watch!
My Fluke meter puts out 3.75 volts in Diode-Check mode so from what I gather that should be enough to test either a 3 or 4 junction diode. When I tested the SAE diode in-circuit with the meter alone I get @ 1.7v. It reads open in the opposite direction. If you think I still should test it with my bench supply then I will do it. I am scared to mess with them too much. They look super fragile and wires are very thin.

Tonight I am going to check all the drivers, trim pot etc again and I will let you know what I find. I believe my high bias voltage is originating at the input board.

 

[elnaldo]

Your tests seem OK, drivers not shorted (always keep in mind that a semiconductor can pass a test with a meter under 3 Volts, but fail in circuit holding 50 or 60V), and bias diode not shorted or open.

If supply voltage is OK, I'd try the circuit without the output transistors again. Check every ground connection to the board is OK. If in doubt, assemble the unit and try everything in it's place.

Check the voltage drop at the diode when powered. Then, that diode should be connector to some transistors. test those too.

Here you have some simple graphs that explain ehats happening in the bias circuit:
https://www.electronics-tutorials.ws/amplifier/class-ab-amplifier.html

Scroll down to "Adjustable Biasing" and to "Class AB Amplifier Diode Biasing". It will help you to understand what voltages to expect, and what to check. We want to know if current is flowing through th ediode, biasing the transistor's bases.

To isolate the problem, you can disconnect the input to the power amp, usually a DC blocking cap that you can remove, or check the DC at the input pin, it should be zero or near zero.

 

[genlabs]

Your tests seem OK, drivers not shorted

I had some time to play around with this today. I pulled the input board and tested every transistor on it. I measured all resistors. Any resistor that measured funky I pulled one leg and measured again. They all tested within 5%.
The caps were replaced recently.
I really cannot find any component that is questionable.

The 5 wire ribbon cable between the boards measure:
5: 60+ Rail Supply
4: 59- Bias line?
3 0 Neg Feedback line
2 59+ Bias line?
1 60- Rail supply

Is there any possibility that lines 4 and 2 are putting out those high numbers because the output transistors are not mounted? I guess what I am getting at is, even a low biased amp will still have some current flowing through the output transistors. Isn't this somehow fed back to biasing circuit to even things out? I just find it funny that the bias voltage wire near the +rail is inverted to negative. The bias line near the minus rail is inverted positive.
I am tempted to put the remaining two good of four transistors back on the board and see if they blow up .

 

[elnaldo]

What's the voltage at the drivers ?
What's the voltage at the output transistors connection tabs? That will tell you if the circuit is OK. Expect +V at the collector tabs, 0.5V at the base tab, 0V at the emitter tab. If you have high voltage at the base or emitter tab, there is still a problem, I'd fix or find out what's happening before installing the output transistors.

I see you are not sure about pins 2 & 4 function since you write a "?". Any open circuit with no current flow because one end is disconnected will show the supply voltage to the circuit when you measure.

Check the DC at the audio output pin, that's the junction of output emitter resistors. Check voltage at the drivers BCE and voltage drop at the driver's emitter resistor.

Build a DBT. it will save your parts if there is a short.