9090DB setting bias didn't go so well

[Blue Shadow]

This is an unrestored unit that has been worked on before. Seems to be in good shape as far as those previous repairs.

I was setting the bias the other day and that didn't go well. I was successful in getting the unit to stay in protection and now I need to troubleshoot the unit.

I'm checking side to side on the driver board and found that one of the TR-05/06 pair is bad now. These B649A seem to be getting replaced with KSA1220AYS, replace both?

The common failed fusible resistors R33/34 are good, one having been replaced in a previous repair, but one is low at 172Ω, the other 181. Any reason to be concerned about these? And if they need to be replaced, do they need to be flameproof resistors?

Other than the two VD diodes, any other components that need to be swapped on this board except the 'lytics?

I have read in another thread that the unit can be tested with the driver board out of the unit as long as the 4-pin connector is attached at the top of the board. Is this true?

Maybe I can get this one back to working.

 

[tnsilver]

My tip is work with a DBT otherwise you're likely to keep frying components. I'd replace both TR-05/06 trannys for pairing and I'd verify all other trannys on that PCB. I would renew/replace all fusibles and EC's as well, like you mentioned.

One special attention deserving aspect are the bias trimmers. These are semi-open design and they tend to get oxidized over time and develop dead spots. They can kill a driver amp in an instant. I would thoroughly clean them with D5 or replace them all together. I'd advise to be very gentle with those.

When the driver amp is out of the unit, it doesn't matter what you do with the 4-pin (input) connector but if the unit stays off protection mode like this - you know for sure the problem is with the driver amp.

What may mess up a driver amp after bias adjustment other then totally shorting the circuit by not taking out the fuse out of the measured fuse holder? A few things... Overshooting the bias current is likely, and so is the trimmers having dead spots. Wrong meter configuration (e.g. meter probes not in series current measurement terminals) and, also very common, measuring one channel while adjusting the other. This happens a lot especially b/c the SM is very erroneous in this regard.

 

[ghazzer]

Is there a better way to measure bias of the 9090db than measuring across the fuse terminals?

I suppose that someone with a working 9090db could measure the bias at the fuse and then measure the voltage across one of the emitter resistors in that channel. That should establish a ratio between the two measurements.

 

[DougBrewster]

The common failed fusible resistors R33/34 are good, one having been replaced in a previous repair, but one is low at 172Ω, the other 181. Any reason to be concerned about these? And if they need to be replaced, do they need to be flameproof resistors?

172 and 181 is inside 5% tolerance for a 180 ohm resistor.

Is there a better way to measure bias of the 9090db than measuring across the fuse terminals?

Not really. Access is difficult for the emitter resistors even with the driver board bracket removed and of course any clumsiness around there will create new issues. What's hard about measuring current across the empty fuse holders?

I suppose that someone with a working 9090db could measure the bias at the fuse and then measure the voltage across one of the emitter resistors in that channel. That should establish a ratio between the two measurements.

The ratio is Ohm's Law..

 

[ghazzer]

The ratio is Ohm's Law..

Thanks Doug! I think that Ohm's law is I=E/R. The ratio is just a ratio.

IF one could reasonably access the emitter resistor(s) and measure the voltage drop to compute the current, then the ratio of that current to the spec bias current can be used to calculate the bias current at the fuse from the voltage across the emitter resistor(s). This assumes that the bias current at the fuse is correct.

 

[Blue Shadow]

Got back to this unit. Check the output transistors and all seem good. Tried the DBT without the 10A left and right channel, measure your bias here fuses in the unit and also did not have the driver amp in the unit. Dim bulb did well at 40w, low lights and such. Went to the 100W bulb and the unit came out of protection. Seems the driver board is all I messed up with the bias adjust attempt. Found all left vs right channel lead to lead resistances the same, essentially, except for TR-05 and TR-06 where were pulled and the TR-05 verified as dead. Guessing I have my shopping list, a couple KSA1220 to replace the good and bad 2SB649A and maybe a couple caps that are bulging.

The new owner of this unit is so frugal he isn't wanting much else done and I haven't told him about this yet, we'll talk tomorrow and I'll see if he wants me to go any further since I'm already inside the unit. Maybe for parts plus some beer, a friend you know. I guess I'll do the VD diodes since I have the 1n4118s in stock. But if he likes the unit and wants it refurbed, it is going to cost him parts and labor if he doesn't spring for some of that now.

 

[Blue Shadow]

Now I'm back to the unit again. After getting involved in a couple other threads on this unit. Still has 49V on the gray wire and I tried all kinds of things, ohmed out both sides of 2624 looking for differences, didn't find any. This after swapping a bad TR05/06. Swapped a few caps on the PS board because once I get this unit fixed I will swap out some of the caps. Spiney isn't wanting a full recap but I did talk him into swapping some of the easy to access caps and the ones that looked bad. Cleaned the VRs on 2624 and then asked Bill if he would print parts of the manual for me so I could trace voltages. Of course, during this investigation, always with a DBT, I've learned another lesson.

Got to checking voltages. Input voltage with the 100 watt bulb in line was 107VAC, getting numbers that seem good, some a bit lower than the diagrams some a bit higher but probably within expected ranges since the unit is not running at designed input voltage. Did this without 2624 in place and the unit would come out of protection. Upon installing the driver board to further test the unit, without the top connector installed, the unit would not come out of protection and after a minute or two the lamp brightness started to cycle. OK, let's start all this part over. Made up the label for the driver board support with bias and offset VR labeled along with the proper Fuse to use and the Channel designations and turned the bias all the way down. Installed the board, without the bracket, with the top connector and the unit came out of protection. I have no idea why other than the bias VRs were not set close at all. Going to search about bias with the emitter resistors to see if I can get in there for that and set it low with the DBT and then try some line voltage and see what happens. Didn't find anything.

Set up to do the bias on the right channel and getting 9.4mA and not much adjustment, less than a quarter turn and I get up to 10mA and the unit goes into protection...while on the DBT. I got no idea. Guess I need to see if the left channel does the same thing. Moving forward...I hope.

 

[LBPete]

The top connector supply ground to part of the board. You need to have that on. Also be sure to have it on the DBT at all times until all issues are sorted out. What is the DC offset of each channel? It's likely going into protection because the DC output voltage is climbing as you turn the bias trimmer.

- Pete

 

[Blue Shadow]

I need to check that, thanks. Got offset to about 10mV on each channel.

Curious why the bias did not change much with the VR movement, but maybe it needs to be moved further. I'll spend some time dialing things in while the unit cruises on the DBT before going live.

 

[Blue Shadow]

Well on the DBT, I see as I decrease, yes, decrease bias there is a change in offset and I make the adjustment. With the ~107VAC into the unit I get about 9.5mA bias and the control is fully CCW or max bias, according to @tnsilver's diagram he put on the F2624 cover.

I see this with both channels and rotation of VR-3 and VR-4 don't do much at all but lower the bias a bit. Haven't gone all the way to minimum. I have the green light and want to keep it. I did go to line power, 122VAC, generally, and the bias shot up to about 12mA and drops from there as I lower the setting. I can't get enough bias. Off to search for that symptom.

Curious what is going on and how to proceed with this unit.

Maybe the unit is not warm enough, will give it some time and see. Meter will be checked but it has been my go to meter that seems to work fine.

 

[Blue Shadow]

While we are here, can we confirm some things?
The Fuses F05 and Fo4, the two in the center of the pic above that are arranged vertically, F05 on the left.
Does F05 feed the left channel and does this use the VR-1 and VR-3 controls? It appears to on my unit, now rocking it old school with a Triplett 60-2 meter hooked up and getting some movement on the meter. If this is true, then the manual is wrong on how to set the bias and offset.

Correspondingly then, F04 would be for the right channel and use VR4 and VR2. Next is to see if they controls are labeled correctly in the above picture.

 

[Blue Shadow]

Getting close. The Triplett allowed adjustment of bias. Used the other meter to check and no reading, fuse. Change that and the digital, both of them read about 65mA but there is no temperature in the heat sinks. But there is heat on the two 270Ω resistors on the tuner power supply board. Going to call it a day on the bias and offset and try to learn the answers to the post just above while I hunt down the overheating resistors. I did recap that board as it has some caps that were in a bad way.

 

[Blue Shadow]

OK on to this tuner power supply issue. The unit plugged into the wall at 122VAC provides 50.some VDC to board F-2619 at R01/02, the two 270Ω voltage dropping resistors. Output is 17.7VDC or so giving about 4.5W per resistor and voltages on this board a few volts higher than the scat shows. I'm guessing this is running correctly and I don't see any adjustment for the incoming voltage which I believe should be 44V.

It appears the 135Ω dropping resistor needs to be able to handle 10W just to be used at its limits. Sansui used two 4W resistors in parallel and I guess I need to figure out what to do about this as the 4.5W through a 5W resistor is getting the board very hot.

Doing some thinking and calculating, the input to the board should be 44V and the output after the two 270Ω resistors should be 14 volts. Knowing that the laws of electricity Sansui used are the same ones as on the web and given the size of the resistors they used, I calculate a current of about 0.22 amps. This info, 44V in, 14V out and 0.22A used requires a 136Ω/6.6W resistor. So Sansui decided to use a pair of 270Ω/5W resistors.

Now that we know the current needed for that board, we can recalculate the resistors needed to get closer to that 14V. It appears that a 165Ohm-8Watt resistor would do the trick without any contingency. But the original design has a bit of extra wattage capacity and to do the same would require closer to 12 watt capacity. Let's look at Mouser. They have some 330Ω/7W resistors that should do nicely.

The other option would be to change the resistor on the main power supply board R-01 that is a 27Ω/3W unit to something that lowers the voltage to 44 as in the design, but I don't know messing with the voltage that is going to the other parts of that circuit.

I can't really go through the scat to determine what is causing the high voltage to the tuner power supply board or how changes in the circuit would affect different parts of the unit. It seems odd that the higher voltage we have today, 122 is causing such a problem, maybe there are some things I need to check.

 

[ghazzer]

Surely some few AKers have been thru this and will give you the benefit of their experience. There are piles of 9090DB threads. I see 500+ for 9090DB & "power supply".

Maybe LBPete will chime in.

 

[restorer-john]

Just a little trick I use for bias settings measured across fuse terminals; old blown 3AG/M205 fuses. Drill out the end cap and put a low value 1/4-1/2 watt resistor inside and solder the ends up. Make up a matching pair, one for each rail.

Then you measure the voltage drop and convert to current. No chance of taking out an output stage because the resistors will smoke up (inside the glass tube- safety first) if there's a problem. No chance of taking out a test probe or meter shunt either.

Make up a selection of values and keep them in your fuse tray for setups.

 

[Blue Shadow]

Having owned Magneplanars for a few years, I have plenty of spent fuses. I kinda collect them in a bowl over there. Told my college buddy when I bought the speakers that he should buy some Buss or Littlefuse stock as he knows I popped fuses in his MG-Is a few times. I need to make a few of those.

What I did was use a removable jumper in place of the fuse with a meter clipped in place on the fuse holder set to mA. Turn the unit on (DBT connected) and remove the jumper to get the reading. Seemed to work fine and got to where the bias is adjusted and the offset is good.

Curious what VR01 on F2619 is used for. Is it to set the 12.5 voltage coming off that board at C?

 

[DougBrewster]

The pot on F2619 is for setting the 12.5V output as you surmise. Note that this setting directly affects the Dolby calibration signal amplitude; you may want to tweak the 12.5V to keep the Dolby calibration correct. The R1/R2 combination is often problematic because of the heat produced and one workaround I've seen in this forum is to install a large 135 ohm power resistor (one of those metal case flange mounted types) on the vertical chassis plate behind F2619; that way the chassis works as a heat sink. The current draw from F2619 is reliably in the 160/175mA range depending whether it's in FM or AM. Whatever resistance value you use for R1/R2 it needs to provide not less than 17V at TR01 collector, not 14V as shown on the schematic. Any lower and the regulator won't have enough headroom to work correctly. About 18-19V is good, go much higher and you may exceed the rating of TR01.

Far as the relatively high voltage you're seeing from the main power supply board is concerned note that if you measure this voltage during a period of grid system load change you will see times when the 120Vac incoming supply increases or dips, sometimes quite significantly. This happens when the kids get home from school and start turning their stuff on for example or in the late evening when folks are turning stuff off. To be sure about the high voltage you appear to have try measuring it at different times of day.

 

[DougBrewster]

"What I did was use a removable jumper in place of the fuse with a meter clipped in place on the fuse holder set to mA. Turn the unit on (DBT connected) and remove the jumper to get the reading. Seemed to work fine and got to where the bias is adjusted and the offset is good."

I noticed this after posting my earlier response. Why use the jumper? It defeats the fuse protection and you have the meter connected anyway with it's own fuse protection.

 

[Blue Shadow]

I have tried testing like this, meter in place of the fuse and lose meter fuses. This was on a different unit that appeared to have a turn on transient above the fuse rating in the meter. That was a 250mA fuse in an auto ranging meter. The 10A setting uses a different probe input.

Thanks for the explanation of f-2619. I'll need to look over the components and move forward on that. Good to see I have a well traveled issue.

 

[tarior]

Find a pair of Pomona 4565 (-0 for black, -2 for red, though color isn't important) grabbers on ebay, make a set of leads with them, then set bias using the emitter resistors. Pulling fuses to set bias with an ammeter is stupid, and dangerous to the circuit under test, IMO.
Also, replacing the bias trimmers on pretty much ANY Sansui amplifier is mandatory.