Marantz 2270 bias won't go below 15mV

I did a mistake in my last post for the value of the resistors of the Vbe multiplier. I used the textbook formula for the Vbe multiplier

Vce = Vbe (Ra + Rb) / Ra

where Ra is the total resistance between the base and emitter of the bias transistor, and Rb is the resistance between the base and collector. This formula is valid only if the bias transistor has a very large hFE.

For low hFE, resistor Rb must be low or you will need a negative resistance for Ra...

For hFE=100, the correct value for Rb is 2.2k and the correct value for Ra is 4.1k, in this amp.

What is going on in your amp is that the range of adjustment of the bias voltage source is very limited and your driver transistors have a Vbe which is too low for your bias voltage source.

Can you measure the actual voltage between the collector and emitter of H760 for different positions of the bias trim pot, for example at 0%, 10%, 50%, 90% and 100% rotation ?
 
Can you measure the actual voltage between the collector and emitter of H760 for different positions of the bias trim pot, for example at 0%, 10%, 50%, 90% and 100% rotation ?

Will do at lunch, thanks for following up. I have a 2285 on the bench with the rear panel off to straighten it out from shipping damage right now but I'll find a flat spot somewhere!
 
Here's the H760 results measuring from E to C:

Bias pot rotation:
0% (CCW) = 1.18Vdc
10% = 1.18Vdc
50% = 1.196Vdc
90% = 1.276Vdc
100% = .706Vdc
 
Thanks for the measurements. As you can see there is a small range of rotation, between 90% and 100% rotation, where the Vce changes rapidly. When the trim pot is at 100% rotation his resistance is ~0 and your Vce is given by the forward voltage drop of H761 (at 13 mA). For rotation higher than 96% (the trim pot has 40 Ohms at this rotation and 40 Ohms X 13 mA = 520 mV) only the diode H761 conducts and the transistor H760 is off.

For rotation less than 96% the voltage across the trim pot becomes sufficient to turn H760 into conduction. At this moment the current from the 13 mA current source is splitted between the H761 and the H760 branches. As the rotation is reduced the current in H761 decreases and his forward voltage drop decreases at the same time. Since the Vbe of H760 is almost constant from 0 mA to 12.5 mA (in fact Vbe increases very slightly, less than the decreases in the forward voltage drop of H761) the total voltage at Vce is reduced as the rotation of the trim pot is reduced.

Your bias voltage could be reduced if you select another diode for H761. Measure the forward voltage drop of the actual H761 with the diode function of your DMM. You will not get 706 mV because your DMM doesn't inject 13 mA into the diode for this test. But you should have no difficulty to find diodes with a lower voltage drop and diodes are inexpensive.

However, I maintain that for best thermal stabilty H760 and H761 must be in thermal contact with H758 and H759 and the bias voltage would be lower if H758 and H759 are warmer than H001 and H002. Before replacing components I would start by relocating these temperature sensing components where they should have been from the start.
 
Thanks for the measurements. As you can see there is a small range of rotation, between 90% and 100% rotation, where the Vce changes rapidly. When the trim pot is at 100% rotation his resistance is ~0 and your Vce is given by the forward voltage drop of H761 (at 13 mA). For rotation higher than 96% (the trim pot has 40 Ohms at this rotation and 40 Ohms X 13 mA = 520 mV) only the diode H761 conducts and the transistor H760 is off.

For rotation less than 96% the voltage across the trim pot becomes sufficient to turn H760 into conduction. At this moment the current from the 13 mA current source is splitted between the H761 and the H760 branches. As the rotation is reduced the current in H761 decreases and his forward voltage drop decreases at the same time. Since the Vbe of H760 is almost constant from 0 mA to 12.5 mA (in fact Vbe increases very slightly, less than the decreases in the forward voltage drop of H761) the total voltage at Vce is reduced as the rotation of the trim pot is reduced.

Your bias voltage could be reduced if you select another diode for H761. Measure the forward voltage drop of the actual H761 with the diode function of your DMM. You will not get 706 mV because your DMM doesn't inject 13 mA into the diode for this test. But you should have no difficulty to find diodes with a lower voltage drop and diodes are inexpensive.

However, I maintain that for best thermal stabilty H760 and H761 must be in thermal contact with H758 and H759 and the bias voltage would be lower if H758 and H759 are warmer than H001 and H002. Before replacing components I would start by relocating these temperature sensing components where they should have been from the start.

Paragraph one makes sense to me. And I was already considering selecting another H761 diode. I have a bench supply and an R substituion box to set the current to whatever it needs to be which is about 13mA as noted. Relocating only the diode to be thermally connected to the top driver transistor is easily done. Moving the TO-220 packaged bias transistor would be a bit more difficult. So perhaps I will try messing with just the diode. A small non-conductive clamp and a spot of heatsink compound for the experiment ought to work. Of course if the main heatsink's temp and the driver's heatsink temp track each other it won't make a difference. This should be interesting!

It may be a day or two until I have the free time to play again so hang in there! Thanks! - Randy
 
I finally got back to this one. What I tried first was to find a diode in the rectifier diode bin with the lowest on voltage. I wanted a rectifier for the heartier body and stout leads. The lowest valued ones I had were old Motorola MR500's which are 3 amp, 50 PIV in value. The original diodes in the 2270 metered out at about 520mV and these were around 485mV. It was enough lower to allow the adjust to be made and is stable. COOL! It worked to repair the other channel too.

First I had to try the commonly available 1N400x type diode on the second channel so selected the lowest value one of those and soldered it in. It didn't work right at all so I took it out and put the old one back in. Still didn't work. Turns out I must have shorted to something because the bias transistor was blown open base to emitter. After studying some specs I subbed in a MJE182 and it was back to working but still couldn't adjust the bias low enough. Back to the 1N400x which yielded an improvement but not quite there. I then put in another MR500 diode and all was well.

I'll have to locate a suitable sub for these diodes as the MR500 is no longer around. Motorola became OnSemi but they don't offer that diode. I'll have to try the 1N540x series as they're similar spec-wise.

I'd like to say thanks again to all who contributed and which led me to the repair. Y'all are the best!! :thmbsp:
 
Fine! I'm happy that you fixed it.

Did you try a "diode connected" transistor? If you connect the base and collector together in a transistor you will have a lower Vbe than you would have with the collector open. This is a very good way to make a low forward voltage drop diode. I am sure you have a transistor bin to experiment...
 
Fine! I'm happy that you fixed it.

Did you try a "diode connected" transistor? If you connect the base and collector together in a transistor you will have a lower Vbe than you would have with the collector open. This is a very good way to make a low forward voltage drop diode. I am sure you have a transistor bin to experiment...

Thanks! I did not try that method but will file it in my book-o-tricks. I wanted to stay with the mounting method used, which is with the diode under a finger of the same clamp that holds the TO-220 style bias transistor to the main amp heatsink. To that end I would endeavor to use a TO-220 style clad in heatshrink for isolating the exposed collector pad on the back. That would fit. I do have a couple in the old bin! ;)
 
Hi, sorry for upping that good old source :) but i have one more question to ask: User "ecluser" talked about to take a 5.6 ohm on R784 and R785 which are carbon types with metrics 3.9 ohm at 1W with 10% tolerance as the service-manual states it. Will a modern type of carbon or metal also work if the tolerance is 5%? Could that lower tolerance harm the circuit? Sorry for that low-level-question, im at the beginning of my hobby-restauration.
 
Last edited:
Hi, sorry for lift that good old source :) but i have one more question to ask: User "ecluser" talked about R784 and R785 which are carbon types with metrics 3,9 ohm at 1W with 10% tolerance as the service-manual states it. Will a modern type of carbon or metal also work if the tolerance is 5%? Could that lower tolerance harm the circuit? Sorry for that low-level-question, im at the beginning of my hobby-restauration.

I believe that any good quality carbon or metal film resistor of at least 1 watt of power dissipation will work. If 10% is specified, the 5% would be even closer to the metric of 3.9 Ohms.
The value (Ohms) of these two resistors can be varied in tandem if the desired bias cannot be achieved, if the bias is too high, increasing the Ohmic value R784 and R758 will deduce it, conversely, if the bias is too low, decreasing the Ohmic value R784 and R758 will increase it. This is providing that there are no other problems and the circuit is operating properly.
I have seen values of zero (a wire) to 5.6 Ohms used for R784 and R785.
As ecluser stated in the post above, the Vbe of the 'bias' transistor has an affect on the bias as well.

At the risk of repeating myself, here are Vbe values I measured for some commonly used 'bias' transistors. They were measured with a Peak Atlas DCA55, the exact value is relative as the base current used for the test will impact the measured value somewhat.

2SD669=0.68V
MJE243G=0.70V
2SD612=0.70V
2SC495=0.72V
KSC3503=0.76V

Tom
 
Agree, to make my situation clear: I have done a total recap on the P750 boards (catcrafter list) plus diodes, drivers and outputs (PhillnAudio Thread BOM). I did relocate of H761 and H760 to the drivers heatsinks. Everything is fine to adjust except the right channel bias wich wont go below 11mV. So i think i will take the chance to exchange the resistors because the H760 is already a new MJE243G.
 
Many people have had good success using MJE150XX TO-220 transistors as drivers in the 2270 and other models.
If I could not get the original TO-66 2SC680 and 2SA566 types I would probably use KSC2073 and it's complement, KSA940 TO-220 devices as substitutes.
(no settings on the scope or curve tracer were changed between photos)

2SC680 (2270 original NPN driver)


KSC2073

MJE15030

YMMV,
Tom

Oops, fixed the pics
 

Attachments

  • 2SC680 TO66 (2270 driver).JPG
    2SC680 TO66 (2270 driver).JPG
    69.8 KB · Views: 26
  • KSC2073.JPG
    KSC2073.JPG
    51 KB · Views: 14
  • MJE15030.JPG
    MJE15030.JPG
    55.3 KB · Views: 16
Last edited:
I can confirm that the TO-66 drivers H758 (2SC680) and H759 (2SA566)
can be replaced with TO-220 drivers MJE15032/MJE15033.
I used them for my 2270 in 2011 (with Tom's guidance here at AK, I may add), and it continues to works great.

These TO-220 drivers fit well into the TO-66 layout and the heat sink can be reused too.
It helps drill out/enlarge the holes for the leads in the old TO-66 Heatsink to make it easier to the fit TO-220 drivers.
 
MJE15032/MJE15033 is what i have put in as drivers. The left channel is doing fine with all those modifications. Just the right channel won't go below 11mV while R763 is full CCW and after 15-20 minutes it raises up to 12.2mV. All other steps i have done before are just fine. Supply is at constantly 35V and "zero"ing the speaker terminals is also fine and stable. Well, for now i will give it a try to test another Diode of Type 1N5401 and check it. If that fails i will try changing the R784/R785 to 5.6 ohm types. If that will also fail i will change the drivers to KSC2073TU/KSA940TU and will see. Did i get that right, that the drivers needs to be matched pairs?
 
Back here and i can confirm that replacing R784 and R785 with a 5.6 Ohm Resistor brought back the possibility to adjust bias with R763 to the correct value (10mV in my case) (which is now pretty exactly in the "middle" of its rotation). And adjusting is now way less fiddly and "sensitive" than before. Again a big thanks to this wonderful community!
 
:needpics:


What can I say?
Are there pics showing to where the bias parts are to be remounted?
Since I don't have one in front of me.
 
Back
Top Bottom