Eight driver board adjustments

Trnsfmr

Well-Known Member
The Eight was a revolutionary model for Sansui. For some reason, they decided to make the differential current, and differential balance, adjustable, on the driver boards.

And then of course, in Sansui's inimitable fashion, they put instructions for adjusting these parameters in the Service Manual, that an EE would find very hard to decipher. It reads like gobbledy gook.

Even though Justin and I figured it out, it's the myriad of ways they tell you to hook up the meter that is confusing. I think this was because there were no digital multi-meters at the time which would display + or - voltages, there was only a scale that started from zero and went up on an analog meter.

Making these adjustments was such a pain, probably their own techs couldn't do it, so they soon came out with the Eight Deluxe, which did away with the adjustments. This is a shame really, as these parameters have more to do with amplifier performance than anything else. When these adjustments are spot on, you get the lowest distortion and best slew rate possible from the design.

Well, now that we have digital multi-meters that will range over + and - voltages so easily, we developed a method for making the adjustments on the Eight very simple. We just rebuilt another one in our shop and went through the procedure again, so this time I took some pictures and thought I would share it here.

How to
First of all understand that there are four things to adjust, one of them being Bias, which everybody knows how to do. The remaining three are:

Differential pair current
The differential pair is fed through VR802 and R804 to a pair of 100 ohm resistor feeding the emitters of TR 801 and TR802. The idea is that no matter what happens signal wise, the total current is suppose to remain constant. That current is 1.5 ma. We can easily measure it by measuring the voltage drop across R804, a 330 ohm resistor. We want to measure 0.500v across R804. VR802 sets this current.

Current balance.
At a no signal condition, we want TR801 and TR802 to have exactly the same current going through them. The actual amount of current is not near as important as the balance. It's much easier to measure the balance instead of the actual current, so we measure the voltage between the two emitters, and set that voltage to 0.000. VR804 sets this balance

Speaker offset
This is no different than we do for any other amp. We like to measure at the headphone jack, one channel at a time to ground. The adjustment for this is VR801. VR801 will only be able to bring the offset in range once you get the other parameters close. Sansui specifies that a resistor (or speaker) be hooked to the output.

SCAN0103.JPG

Each one of these adjustments affects the others. The way we deal with this is to have a meter for each parameter. We hook them all up at the same time, and just go back and forth with adjustments until they are spot on, with everything warmed up, of course. If you have only one meter, I would really recommend that you go out and buy a couple more. Harbor freight meters are fine, we buy lots of them, and they are cheap. The little one is really cheap, and there's no excuse not to have several.

I blow up meters all the time, much to Justin's amusement. We probably buy most of the meters that our local Harbor Freight sells. We're a good customer. HA!

First let's look at setup and then the adjustments.

012.JPG

In the first picture you can see where the meters are hooked into the driver board. We use the little mini-grabbers, they work great. The red and black leads from one meter are connected to R804. The yellow and white leads to another meter are connected to the emitters of TR801 and TR802.

014.JPG

In the second picture you can see 5 meters all hooked up at the same time. The first meter on the lower left is hooked to R804 (2v. range). The second meter, right next to it, is hooked to the emitters (200mv range). The third digital meter is on the shelf, hooked to Left headphone output (200mv range). The two big Tripplett analog meters are reading bias for the left and right channels. An analog meter is so much easier to read for some measurements.

We like to meter everything when we are adjusting an amp. If you can see everything, then there's much less chance of something happening, and blowing a pair of output transistors in a few seconds. The Trippletts are fused internally, so if bias goes over 1 amp, it blows. We also have 1 amp fuses in line to our bench speakers. After years of blowing things up, you get wary.

We put 20 turn adjustable pots in, because the old ones were so scratchy that they would blow fuses in the meters and speaker lines. It's a 47 year old amp, after all.

Step by step
So... if you changed transistors or other repairs to the driver board, set VR801, 802 and 804 to midrange. Otherwise just start from where they are. VR803 is bias and will be the first thing to adjust.

1. Turn amp on and immediately set VR803 for bias. Factory spec is 8-10ma.
2. Set VR802 for 0.500v. on the appropriate meter.
3. Set VR804 for 0.00 on the appropriate meter
4. Set VR801 for as low a reading as you can get.

5. Let it warm up and do 2 - 4 several times until it's stable. Bias will be pretty stable and mostly change because of warm up. The other three really affect each other and also change with warmup.

The factory instructions talk about test points 1-3. Forget about them. The above is all you need to do.

We change the differential pair as a matter of course on every driver board we work on. Usually KSA992's. If you have trouble getting measurements in range, change them.

That's it. A long explanation, but the drawing pretty much explains it. I had to draw it out, because I have never found a good schematic for the Eight that I could print out or capture. Hope this helps.
 
Last edited:
Jim - This is huge. Thanks for cracking this nut.I have struggled with that adjustment description as well. The manual addresses vr801/804 to be for offset - I assumed one was coarse the other fine - which would make sense on an amp so obsesses with adjustment that they would even let your adjust the differential current..Then as you move on to the next adjustment - "current adjustment on differential amplifier" - you are asked to adjust vr802 until the last step (7) where vr801 is called for ?. This never made any sense to me as you just set it during the offset adjustment, which would now be messed up. I looked at the schematics and saw vr802 connected to the emitter resistors of the differentials TR801/02. Therefore I assumed the call for vr801 was another sansui typo

Now here comes the strange thing - ignorance is bliss as they say - so I decided I was right and the manual wrong and then stuck with vr802 as the only pot to be turned during the "differential adjustment", despite before mentioned step 7. Yet, turning vr802 at this point gave me the zero called for (of course the probes would be on the appropriate TP's depending on the probing from step 5,6.) As you describe I went through the steps a couple of times and subsequently got the most perfect offset I've ever seen - a zero so steady on the DMM that I thought something was wrong. The amp sounded very good, yet it must have been out of adjustment :crazy: - worse part is I have advised others to this procedure :(, because it seemingly worked, whereas following the directions gave me confusing readings.

On a side note what do you think of Eight amp ?
 
Hi Tom,
I forgot to mention something that is important. On my scribbly drawing I show VR802 and VR804 as 5k ohms. Not so. They are 3k trimpots from the factory. When I replaced them with 20 turn trimmers, only 5k's were available so that's what I used.

This doesn't make any difference for VR802, but it does make a difference with VR804. To make everything work as it is supposed to, I replaced the 3.9k resistors right above VR804 with 2.9k. (I fixed this in the drawing in above post) Not a standard value, but I have huge racks of 1% resistors. That way, the total resistance was the same for the collector path to B- voltage. It needs to be that for VR801 to come into range so you can adjust offset.

It might be interesting for you to see where your settings are, metering as described above. I'd be curious to hear. I admire you for getting through it at all. I pretty much gave up understanding it and just figured it out from what I knew was needed from amplifier circuit design. Justin understood the instructions and set an amp or two using the instructions. This way is far better and faster though.

And yes, I've been playing the Eight in the shop for the last two days. Every time I play with one of these, I am amazed. It's basically the same sound as the AU999, at least to my ears and... a lot of power with very good low end. And that makes sense, because right after the Eight, came the 777 and finally the 999.

I just love the way it looks, too. especially with cabinet and front panel stripped off.

I was hoping you'd see this and think it worthwhile... your avatar and all. Thanks.
 
Last edited:
Every time I fix and adjust one of these I have to relearn the procedure and decipher the SM and look up the tricks on AK.

Thanks for this, except now I want to dig out my eight again and check the adjustments....again.

So, do you set the bias the SM way with current readings or the easier way with voltage drop across the emitter resister?
 
Last edited:
"So, do you set the bias the SM way with current readings or the easier way with voltage drop across the emitter resister?"

Hey, sorry for not being clear. We set the bias by pulling the fuses and putting the meter in circuit, the standard way. Measuring the drop across the resistor should work, but I worry about the resistor being off a little.
 
great info, will come in very handy when I get to servicing my Eight

Re the bias setting via the emitter resistors, do you set this as per the service manual value, or a little lower (as the service manual value measuring current at the fuse also includes the driver current). I have made up a simple method of measuring the voltage at the fuse



in place in an AU-7500 ready for measuring



a 1 ohm resistor soldered across a blown fuse or fuse lamp, fit in place of the fuse and measure the voltage (1mv = 1ma). I have compared this method to setting bias via the emitter resistor voltage method in several different Sansui amps, and typically you end up with anywhere from 10% to 20% higher bias current than that specified when setting via the emitter resistor.

Cheers

John
 
Hi Jim - I sold my Eight - you'll get a kick out of that story here it is - otherwise I'll be in full swing re-adjusting, measuring and reporting back.

@ skippy124 - great idea with the modded fuse and its 1 ohm = 1mV = 1mA resistor.
 
Squawk 7700 and extinguish. You a pilot?

Yes - commercial background from Denmark, now just putting around on a Private license. In a few month a will have my own aircraft and resume flight instruction as a side business...there will most definitely be a Sui in my hangar :music:
 
Wow, a real pilot, that's very cool. I've been building experimentals since 2000. I'm currently flying a motor glider type, from a Ukraine kit, an Aeroprakt A20. Light sport license, no medical required. I have an old Sony receiver in the hanger, that someone gave me. I can't keep Sansui units for myself, some one always wants to buy it, and my son Justin will say, "Just sell it Dad, we can always find another one".

The two hobbies really go together for me, somehow. Flying is music for the soul. And so I listen to music when I can't be flying.

My signature on a couple news groups says... "Any landing you can walk away from is a good landing. An excellent landing is when you can fly the same plane again, without having to fix it first."
 
My old thread was created because, as noted in this thread, the damn manual is puzzlement. I finally got adjustments done, but with no real understanding of what i was doing. Having read this thread, I still don't understand. :rflmao: The Eight has been my bedroom system for 10 years or more and in two different houses. It is now on its way out because it is letting out some really loud pops at random times during operation. I'm going to have to pull it out of rotation and have it rehabbed. Thanks for this thread, as it has given me answers to things I wonder about, but I still have a lot to wonder about. :D
 
Well, it is pretty simple really. The essence of the modern amplifier is contained within the circuitry around the first few transistors, especially the first two, which is called the differential pair. Any differential pair is designed for a certain current level, which on Sansuis, ranges from about 1.3 - 2.7 ma for different models. This pair of transistors amplifies the input signal and compares it to the final output that drives the speakers

This, of course, is called negative feedback, and for it to function at its best, the current in these two transistors should be the same at a no signal condition, or "balanced". Most other things being equal the amplifier will make its lowest distortion at the designed current level. Distortion rises precipitously when the balance goes away. A 10% - 20% imbalance (not uncommon) in current will usually raise distortion 4X or more.

The other thing that happens with imbalance is the positive going part of the output will have a different slew rate than the negative part of the output. So best slew rate will be at a balanced condition also. Even 1% resistors on a production run will not be good enough to insure balance within a few percent, due to other parts tolerances. The only way to achieve good balance is to have some means of setting it, like in the Eight.

Can you hear the difference? I think you can. A properly set up Eight sounds very good indeed.
 
Last edited:
Well, it is pretty simple really. The essence of the modern amplifier is contained within the circuitry around the first few transistors, especially the first two, which is called the differential pair. Any differential pair is designed for a certain current level, which on Sansuis, runs from about 1.3 - 2.7 ma for different models. This pair of transistors amplifies the input signal and compares it to the final output that drives the speakers

This, of course, is called negative feedback, and for it to function at its best, the current in these two transistors should be the same at a no signal condition, or "balanced". Most other things being equal the amplifier will make its lowest distortion at the designed current level. Distortion rises precipitously when the balance goes away. A 10% imbalance in current will usually raise distortion 4X or more.

The other thing that happens with imbalance is the positive going part of the output will have a different slew rate than the negative part of the output. So best slew rate will be at a balanced condition also. Even 1% resistors on a production run will not be good enough to insure balance within a few percent, due to other parts tolerances. The only way to achieve good balance is to have some means of setting it, like in the Eight.

Can you hear the difference? I think you can. A properly set up Eight sounds very good indeed.

I will say, that this description sure helps me on what is going on.

Thank you Trnsfmr.

George.
 
Hello,
I have had Eight few months now. Bought it from the first owner's son. It was sold "as is", full working order, original state. It seemed that everything was working fine and it sounded absolutely fantastic! I have been thinking of full restoration in the near future.
Now I just wanted to check that adjustments were all in order with the original components. I was reading service manual and it really wasn't easy to follow. So I found this thread and it seemed to be simple enough even a non-professional like me to follow. BUT, seems that I messed up something.
First I adjusted bias. Took off F001 and F003 and put meter onto those holders. I've multiple meters and only the analog one showed values for this one. Right channel was pretty much in the spec value and with a little adjustment also left ch was set to 9 ma.
Then I started steps (with right channel):
2. Set VR802 for 0.500v. on the appropriate meter.
3. Set VR804 for 0.00 on the appropriate meter
4. Set VR801 for as low a reading as you can get.
I think it was on step 4 that I had some problems with the digital meter on it and I changed the settings of that one. Then I noticed that my analog bias meter 9 ma value jumped high up. Now the bias is stucked more than 500 ma and I can't get it down to spec values anymore by adjusting VR803, nor vr801-4.
It seems that also the left channel bias went high after this messing up. It is possible that my meter connections in the headphone plug was messed somehow.
Then I raised my hands and haven't done anything else. Just turned the amp off.
Suggestions what did I burn or messed up and what would be the next steps to fix the issues, thanks!

Toni
 
Hello,
I have had Eight few months now. Bought it from the first owner's son. It was sold "as is", full working order, original state. It seemed that everything was working fine and it sounded absolutely fantastic! I have been thinking of full restoration in the near future.
Now I just wanted to check that adjustments were all in order with the original components. I was reading service manual and it really wasn't easy to follow. So I found this thread and it seemed to be simple enough even a non-professional like me to follow. BUT, seems that I messed up something.
First I adjusted bias. Took off F001 and F003 and put meter onto those holders. I've multiple meters and only the analog one showed values for this one. Right channel was pretty much in the spec value and with a little adjustment also left ch was set to 9 ma.
Then I started steps (with right channel):
2. Set VR802 for 0.500v. on the appropriate meter.
3. Set VR804 for 0.00 on the appropriate meter
4. Set VR801 for as low a reading as you can get.
I think it was on step 4 that I had some problems with the digital meter on it and I changed the settings of that one. Then I noticed that my analog bias meter 9 ma value jumped high up. Now the bias is stucked more than 500 ma and I can't get it down to spec values anymore by adjusting VR803, nor vr801-4.
It seems that also the left channel bias went high after this messing up. It is possible that my meter connections in the headphone plug was messed somehow.
Then I raised my hands and haven't done anything else. Just turned the amp off.
Suggestions what did I burn or messed up and what would be the next steps to fix the issues, thanks!

Toni
Ok, so I somehow damaged my old analog meter doing that step 4. Now, with new meter readings are what they should be and I got my Eight adjusted into exactly recommended values!
Great to have this kind of instructions threads!!
 
The Eight was a revolutionary model for Sansui. For some reason, they decided to make the differential current, and differential balance, adjustable, on the driver boards.

And then of course, in Sansui's inimitable fashion, they put instructions for adjusting these parameters in the Service Manual, that an EE would find very hard to decipher. It reads like gobbledy gook.

Even though Justin and I figured it out, it's the myriad of ways they tell you to hook up the meter that is confusing. I think this was because there were no digital multi-meters at the time which would display + or - voltages, there was only a scale that started from zero and went up on an analog meter.

Making these adjustments was such a pain, probably their own techs couldn't do it, so they soon came out with the Eight Deluxe, which did away with the adjustments. This is a shame really, as these parameters have more to do with amplifier performance than anything else. When these adjustments are spot on, you get the lowest distortion and best slew rate possible from the design.

Well, now that we have digital multi-meters that will range over + and - voltages so easily, we developed a method for making the adjustments on the Eight very simple. We just rebuilt another one in our shop and went through the procedure again, so this time I took some pictures and thought I would share it here.

How to
First of all understand that there are four things to adjust, one of them being Bias, which everybody knows how to do. The remaining three are:

Differential pair current
The differential pair is fed through VR802 and R804 to a pair of 100 ohm resistor feeding the emitters of TR 801 and TR802. The idea is that no matter what happens signal wise, the total current is suppose to remain constant. That current is 1.5 ma. We can easily measure it by measuring the voltage drop across R804, a 330 ohm resistor. We want to measure 0.500v across R804. VR802 sets this current.

Current balance.
At a no signal condition, we want TR801 and TR802 to have exactly the same current going through them. The actual amount of current is not near as important as the balance. It's much easier to measure the balance instead of the actual current, so we measure the voltage between the two emitters, and set that voltage to 0.000. VR804 sets this balance

Speaker offset
This is no different than we do for any other amp. We like to measure at the headphone jack, one channel at a time to ground. The adjustment for this is VR801. VR801 will only be able to bring the offset in range once you get the other parameters close. Sansui specifies that a resistor (or speaker) be hooked to the output.

View attachment 703389

Each one of these adjustments affects the others. The way we deal with this is to have a meter for each parameter. We hook them all up at the same time, and just go back and forth with adjustments until they are spot on, with everything warmed up, of course. If you have only one meter, I would really recommend that you go out and buy a couple more. Harbor freight meters are fine, we buy lots of them, and they are cheap. The little one is really cheap, and there's no excuse not to have several.

I blow up meters all the time, much to Justin's amusement. We probably buy most of the meters that our local Harbor Freight sells. We're a good customer. HA!

First let's look at setup and then the adjustments.

View attachment 702876

In the first picture you can see where the meters are hooked into the driver board. We use the little mini-grabbers, they work great. The red and black leads from one meter are connected to R804. The yellow and white leads to another meter are connected to the emitters of TR801 and TR802.

View attachment 702877

In the second picture you can see 5 meters all hooked up at the same time. The first meter on the lower left is hooked to R804 (2v. range). The second meter, right next to it, is hooked to the emitters (200mv range). The third digital meter is on the shelf, hooked to Left headphone output (200mv range). The two big Tripplett analog meters are reading bias for the left and right channels. An analog meter is so much easier to read for some measurements.

We like to meter everything when we are adjusting an amp. If you can see everything, then there's much less chance of something happening, and blowing a pair of output transistors in a few seconds. The Trippletts are fused internally, so if bias goes over 1 amp, it blows. We also have 1 amp fuses in line to our bench speakers. After years of blowing things up, you get wary.

We put 20 turn adjustable pots in, because the old ones were so scratchy that they would blow fuses in the meters and speaker lines. It's a 47 year old amp, after all.

Step by step
So... if you changed transistors or other repairs to the driver board, set VR801, 802 and 804 to midrange. Otherwise just start from where they are. VR803 is bias and will be the first thing to adjust.

1. Turn amp on and immediately set VR803 for bias. Factory spec is 8-10ma.
2. Set VR802 for 0.500v. on the appropriate meter.
3. Set VR804 for 0.00 on the appropriate meter
4. Set VR801 for as low a reading as you can get.

5. Let it warm up and do 2 - 4 several times until it's stable. Bias will be pretty stable and mostly change because of warm up. The other three really affect each other and also change with warmup.

The factory instructions talk about test points 1-3. Forget about them. The above is all you need to do.

We change the differential pair as a matter of course on every driver board we work on. Usually KSA992's. If you have trouble getting measurements in range, change them.

That's it. A long explanation, but the drawing pretty much explains it. I had to draw it out, because I have never found a good schematic for the Eight that I could print out or capture. Hope this helps.



So stupid question - where are the hook-up points to measure the bias - I see the adjust os done on VR803 but am missing where to hook up the multi-meter.
thanks in advance!
 
Yes, the SM is confusing, and you have to be mindful of the polarity when attaching the multimeter. The SM shows the four fuses F001-4 in the picture of the bottom of the chassis. F001 and F003 are the top two fuses as you look at the underside of the chassis. Here are my drawings of the fuse locations and polarity that I made at the time as well as the location of the driver board trimmers. “Back” indicates the back of the receiver. With that picture, follow the SM procedure on page 16. I found that if you try to set the current with the receiver on a DBT, you will not get the spec values. I was only able to get that on full line voltage.

index.php
 
Yes, the SM is confusing, and you have to be mindful of the polarity when attaching the multimeter. The SM shows the four fuses F001-4 in the picture of the bottom of the chassis. F001 and F003 are the top two fuses as you look at the underside of the chassis. Here are my drawings of the fuse locations and polarity that I made at the time as well as the location of the driver board trimmers. “Back” indicates the back of the receiver. With that picture, follow the SM procedure on page 16. I found that if you try to set the current with the receiver on a DBT, you will not get the spec values. I was only able to get that on full line voltage.

index.php


Thanks Keith for this information. This has been very helpful and my first foray into adjusting these receivers. So I was pretty comfortable attempting this today but have missed something:

Removed F1 & F3 (there being two fuses per channel, I only removed the top two F1 & F3, left presumably F2 & F4 in place).
Turned receiver on.
Set multimeter to mA reading.
Probed F1 with the correct +/- polarity. Left channel.
Read 4.2mA

Did same for the right channel: 4.2mA

Left channel emitter resistor R832 - measured 525mV across the installed resistor.

VR803 was set mid-way but this difference obviously too much to correct with the trim pot.
I'm missing something here. Appreciate any help - thanks in advance !!
 
Back
Top Bottom