I'm confused, are the devices you were looking at in the context of posts 9 and 10 not BD137 and BD138
Yep. The board has two fuses, one per output stage. I'm using 250mA fuses right now.
In my amp there were BD175 BD178 soldered in. But as I said, the board has printed on it self the words BD131 BD132 around the output transistors solder terminals.
I'm not using quadro effect, and I doubt I would ever turn the vol knob past 12:00. So let's hope everything's fine.
The quadro effect is weird. I've been tracing it; the right front output also goes to the rear left and rear right outputs. In quadro mode, a 5.6r resistor is put in parallel from output to ground and the circuit is closed.
In my amp there were BD175 BD178 soldered in. But as I said, the board has printed on it self the words BD131 BD132 around the output transistors solder terminals.
I'm not using quadro effect, and I doubt I would ever turn the vol knob past 12:00. So let's hope everything's fine.
The quadro effect is weird. I've been tracing it; the right front output also goes to the rear left and rear right outputs. In quadro mode, a 5.6r resistor is put in parallel from output to ground and the circuit is closed.
Regarding the hfe selection mentioned by OldDIY. Is it a 152hfe & 188hfe a good match? or should it be even closer?
Just some historical context - I remember quadraphonic projects based on a stereo set up and even built one of these up. It didn't do a lot - mostly background ambience.
I still have an unused Motorola logic chipset for the SQ matrix using 4 separate channels.
This project used a regulated power supply of 35 V. With this the 1% distortion power output was 15 Watts for two channels and 12.5 Watts for four.
The designation of the amplifier modules was SI-1010-Y
For details see https://steampoweredradio.com/pdf/sanken/sanken hybrid audio amps catalog circa 1978.pdf
These modules were compact but I could not find the supplier.
Hi nonost,
In the spirit of 'continued learning', a couple more bits ..
For the same resistance value in the same circuit, a 1/2 watt resistor or a 1 or 2 watt resistor all dissipate the same amount of heat. A lower power-rated part might reach a higher temperature, but the same calories or BTU's will be shed. A part with too-low of a rating may stretch (to a higher resistance value) from excessive heating, or even fail open. Whether that is 'catastrophic' depends on the design and the part's location; a component that determines bias of an output stage is a good example of a place where you don't want want either stretching or failure open. That will often take out the output devices well before the fuse blows.
There was also a lot of gear from Japan in the 1970s that used flameproof/fusible resistors. The idea was to run it close enough to its rating that it would pop if anything went wrong, saving more expensive parts in the process. I don't see any sign of that notion here.
I would recommend like-rated replacement parts until you acquire more experience.
This circuit is one that underwent intensive study in an effort to eliminate absolutely every possible component to cut costs. When that's been done, you have a design that works with the original transistors, but struggles with replacements -- even 'exact' replacements. Each time the manufacturer received a new shipment of transistors, the passive component values would be adjusted to accommodate them. It is simply expecting too much of such a design that just popping in some new transistors will work straight away -- without any resistor values being adjusted.
If the 31,75uA given by 970k (trimpot max + 470k) is still too much, just sub out the 470k with a 680k or 820k, and move on with your life. The object is to center the output node mid-supply by adjusting the Base current of the input transistor. There's no need to worry about the other voltages -- they're fine.
Cheers
In the spirit of 'continued learning', a couple more bits ..
For the same resistance value in the same circuit, a 1/2 watt resistor or a 1 or 2 watt resistor all dissipate the same amount of heat. A lower power-rated part might reach a higher temperature, but the same calories or BTU's will be shed. A part with too-low of a rating may stretch (to a higher resistance value) from excessive heating, or even fail open. Whether that is 'catastrophic' depends on the design and the part's location; a component that determines bias of an output stage is a good example of a place where you don't want want either stretching or failure open. That will often take out the output devices well before the fuse blows.
There was also a lot of gear from Japan in the 1970s that used flameproof/fusible resistors. The idea was to run it close enough to its rating that it would pop if anything went wrong, saving more expensive parts in the process. I don't see any sign of that notion here.
I would recommend like-rated replacement parts until you acquire more experience.
This circuit is one that underwent intensive study in an effort to eliminate absolutely every possible component to cut costs. When that's been done, you have a design that works with the original transistors, but struggles with replacements -- even 'exact' replacements. Each time the manufacturer received a new shipment of transistors, the passive component values would be adjusted to accommodate them. It is simply expecting too much of such a design that just popping in some new transistors will work straight away -- without any resistor values being adjusted.
If the 31,75uA given by 970k (trimpot max + 470k) is still too much, just sub out the 470k with a 680k or 820k, and move on with your life. The object is to center the output node mid-supply by adjusting the Base current of the input transistor. There's no need to worry about the other voltages -- they're fine.
Cheers
I'm back. Main problem solved after installing new trimmers. Sealed new ones.
After some listening I've found two issues.
The first one is some kind of distortion at quiet passages. At low volume. For example, at the intro of Radiohead - Glass Eyes. I've bypassed the preamp, so it's not the problem.
The 5.6r resistor is 6r. I got a new one, but it also reads 6r...I don't know if that difference is problem. Transistors aren't the problem; I've tried different couples from different models (bd131 bd132, bd175 bd178) with no luck.
Does it need more than one diode at bias?
The second one is an annoying noise at the background. It's not hum. It's almost continuous, like "pt pt pt", very close. Changed the big 4700uf again and didn't help. It isn't loud, but audible enough at low volume. But it isn't hum.
Any hint?
After some listening I've found two issues.
The first one is some kind of distortion at quiet passages. At low volume. For example, at the intro of Radiohead - Glass Eyes. I've bypassed the preamp, so it's not the problem.
The 5.6r resistor is 6r. I got a new one, but it also reads 6r...I don't know if that difference is problem. Transistors aren't the problem; I've tried different couples from different models (bd131 bd132, bd175 bd178) with no luck.
Does it need more than one diode at bias?
The second one is an annoying noise at the background. It's not hum. It's almost continuous, like "pt pt pt", very close. Changed the big 4700uf again and didn't help. It isn't loud, but audible enough at low volume. But it isn't hum.
Any hint?
"pt pt pt" may be the result of self-stimulation or may be 1-2 BJT noise.
To increase the quiescent current, you need to choose a resistor instead of 5r6 (increase). Control the increase in the current consumed by the amplifier within + 10-20mA.
To increase the quiescent current, you need to choose a resistor instead of 5r6 (increase). Control the increase in the current consumed by the amplifier within + 10-20mA.
Last edited:
Logically, the meter you are using doesn't have sufficient precision or range settings to read low resistance values below 1R. Did you check the available resistance range settings? If you are using a hand-held multimeter, it may even be a low battery problem.
Last edited:
The motorboating issue has disappeared. I don't know exactly what it was; I changed the 22pf for a new one, reflow solder joints and some other stuff.
The remaining problem is the distortion I mentioned in the previous comment. I'm going to try other values in place of the 5.6r. What's the power rating needed for this resistor? I only have low wattage resistor to do the tests. Then I will install a proper one.
The right channel distorts a bit more. The only difference is the diode used. In the left channel there's the original with a Vf of 0.66v and in the right channel a 1n4001 with 0.55v (I fried the original time ago). I don't know if that might be the cause. If so, can I add another diode in series? A low vf one, like 1n5817.
My tester is precise enough. It even reads the leads' resistance. Which is around 0.5r.
The remaining problem is the distortion I mentioned in the previous comment. I'm going to try other values in place of the 5.6r. What's the power rating needed for this resistor? I only have low wattage resistor to do the tests. Then I will install a proper one.
The right channel distorts a bit more. The only difference is the diode used. In the left channel there's the original with a Vf of 0.66v and in the right channel a 1n4001 with 0.55v (I fried the original time ago). I don't know if that might be the cause. If so, can I add another diode in series? A low vf one, like 1n5817.
My tester is precise enough. It even reads the leads' resistance. Which is around 0.5r.
Last edited:
Sure, depending on the length and cross-sectional area of the copper flex in the leads, there is always some resistance to compensate for, whenever that 0.5R resistance is a significant part of the reading. Effectively, any resistance measurement below 100R should be compensated by subtracting the measured total lead resistance.
So with multimeters, a reading of 6.0R for a 5% tolerance resistor of 5.6R is likely correct but 5.6R would be wrong. Generally, there is no automatic lead compensation because unspecified leads may be fitted or even direct connections to the sockets (which omit any leads) may be used. Unavoidably, users must make the appropriate corrections to the readings themselves. In forum posts, it makes sense to perform the corrections first, so we understand what is written to be true.
So with multimeters, a reading of 6.0R for a 5% tolerance resistor of 5.6R is likely correct but 5.6R would be wrong. Generally, there is no automatic lead compensation because unspecified leads may be fitted or even direct connections to the sockets (which omit any leads) may be used. Unavoidably, users must make the appropriate corrections to the readings themselves. In forum posts, it makes sense to perform the corrections first, so we understand what is written to be true.
Last edited:
https://lh3.googleusercontent.com/p...B8BJLpnyicWKbqh1v4CfX13ZFJROYcRpyoyzN5ofuW4V4
can I add another diode in series? A low vf one, like 1n5817 - Yes!
Your task is to get true bias between the bases of the output transistors within 1.1-1.2v or a current of 10-20mA in 0.5r. Avoid excessive current. This will heat up the transistors and increase the hum to 100 (120) Hz.
1. You can increase the resistance of the 5r6 resistor to 15-20 ohms.
2. Install a diode instead of 5r6 (observe polarity from + to-).
Select diodes according to the total voltage drop.
3. Build the Ube circuit: in place of the diodes, the divider resistors. You can install pot 1k. Add transistor, connect with resistors (base to junction point).
Last edited: