Scott R75s rebuild questions

[Scott75s1974]

Ok, I fired up the DBT and proceeded with the testing plan.
Initial run, all boards out. (1) Tested PS, filter caps and rails. Worked as planned (initial somewhat bright bulb, which quickly faded as caps charged. Then dark.)
Proceeded to full power through an isolation transformer. (2) Added and checked regulator board. (3) Added driver boards and checked them. (DBT check on each new step.)
AM, FM, IF and PreAmp boards were not in chassis for these tests. All looks good. I'd appreciate comments on these numbers. Especially Bias and Offset adjustments at end.

Meter is Extech MN16A. Which seems to be accurate. (How do I really know though?)

Measurements:
(Specifications) Actual
Secondary Voltage (55 VAC) 61 VAC ???
DC off of Rectifier 81 VDC

B+ Rail (+40 +/- 4v) +40.4 VDC
B- Rail (-40 +/- 4v) -40.4 VDC

Regulated Voltage
(+28 +/-2.8v) +26.9VDC
(+13 +/-1.3v) +12.2

Measurements

Bias (Spec .2mA .5mV -- Gives both specs)
Before/After Bias Adjust
L .7 mA /.2 mA
R. 20 mA/.2 mA

*** After adjust both read .000 VDC Spec is .5mV +/- 0*** Is this ok?

I was surprised that I had to turn the trimpots in opposite directions on the two boards to lower the reading. The trimpots are single turn, but different models. Maybe resistor is on different sides. I'll check that.) Even though they were single turn they had a pretty fine adjustment. I guess that's the design value of 275 ohm trimpots... which are not at all easy to find replacements for. There was no evidence of jumping as I adjusted so I guess they stay. Since they worked, I think I'll skip the Faderlube on them.

Output Offset (Spec +/- 1.2mA, +/-25mV)
Before Bias Adjust [No separate Offset adjustment trimpot]
L -1.2 mA / .014 VDC
R -2.7 mA / .007 VDC

After Bias Adjust
L .000 mA / .009 VDC (Can drift up to .019)
R .000 mA / .009 VDC (Drifts down to .005)

(L can drift up to .019. Putting a cooling finger on T1 brings it down. Cooling its partner T2 moves it up. Fun to watch differential gain matching in action.)

That seems very good to me. Like maybe too good?

 

[Scott75s1974]

The Service Manual also gives an Ohmmeter test for the Output Transistors.
These did not really work. The reading kept increasing for 3 out of 4. Which is the sort of thing I expect for in-circuit testing.
Yet this is the first test on the extensive "Audio Test" section of the SM.

It specifies a Triplett 630A ... so maybe there is a Ohm/Volt assumption there that doesn't hold for my meter?

All supposed to read 1.5kOhms +/1 150 Ohms
---> is my ASCII attempt to describe the meter moving
L
PNP 1.4k
NPN 4k--->+

R
PNP 1k --> 2k....+
NPN 1.5k -->+

I wouldn't sweat this except that it's listed in the SM.

 

[Scott75s1974]

I'd welcome advice for the next steps. Here is my plan:

1. Work through the driver boards with meter checking voltage point readings.
2. Test the driver boards.

I don't want to involve the Preamp until I replace the faulty resistor on that. And since I don't have a 680k on hand, that will have to wait for the big order.

The back panel has RCA jumpers (solid conductors shaped like a U) between Accessory Out and In. From the schematics these are located between the Preamp and the Tone boards.

So, I assume, that I can put an audio source (like a smartphone) into the Accessory In side and be feeding the Drivers directly.
Run them hard and watch what happens.

How does that sound?

Thanks again for all the help.

 

[Scott75s1974]

Results of driver test into two sets of speakers. Heathkit AS-2A's and Optimus Pro 7's (separately).

Noise: Full volume, no input -- slight hiss from speakers -- audible from about a foot away from Opt7's. How is that?

Voltage spike at turn on and turn off. Serious thump from speakers. Voltmeter reads a spike of about 1.6 V but I suspect its passes too quickly to get a read. There is no speaker protection relay in this. Is this thump a symptom of a problem with components; or merely a fact of design w/o protection relay?

I'm feeding the audio jack output of an iPhone directly into the drivers. Set the iPhone about half volume. It can handle full volume as well. And that moves some air. (Unrelated? but true fun fact: while I was doing this a transmission line fuse blew and took out half of the houses on my block.)

Performance. I played the Suite from Planet Earth II, which has a good dynamic and tonal range.
It was great to hear some music in the bass range that I've never heard before coming alive on the AS-2A's. Serious music down low. Great imaging.
These speakers are not great in the higher ranges. Both speakers did fine though on discrete or low density high notes. Great imaging in bass and high with discrete notes.

It was the midrange and upper end is where problems occurred. This piece has a big choral component and at peak points things just got very noisy. There's a lot going on in the midrange (?) -- Sopranos, Altos, trumpets, violins. It had all the power needed to play the simple bass music or simple high music well. But when everything was happening at once and loudly, it got bad. I've notice this on other systems at this particular point as well (it works well on mid-market Klipsch earbuds). There's a certain noisiness. I wonder if this is the sort of crowded audio space is where amps get truly tested. How can I distinguish whether this is clipping or distortion?

Is this the sort of problem that a recap might help? (If it's clipping, the PS recap should help - right?) Or is it inherent in the architecture? (As I begin to think about what's happening to the signal at that point I realize why transistor speed might matter in amps even though audio frequencies are nowhere near transistor rated frequencies.)

 

[Leestereo]

Measurements:
(Specifications) Actual
Secondary Voltage (55 VAC) 61 VAC ???
DC off of Rectifier 81 VDC

B+ Rail (+40 +/- 4v) +40.4 VDC
B- Rail (-40 +/- 4v) -40.4 VDC

Regulated Voltage
(+28 +/-2.8v) +26.9VDC
(+13 +/-1.3v) +12.2

The transformer secondary voltage does seem high (~10%), but since the rail voltages and the regulated voltages are within spec, nothing to worry about.

Measurements

Bias (Spec .2mA .5mV -- Gives both specs)
Before/After Bias Adjust
L .7 mA /.2 mA
R. 20 mA/.2 mA

*** After adjust both read .000 VDC Spec is .5mV +/- 0*** Is this ok?

These specs are odd, the bias current is often 10-30mA, e.g., a bias of 20mA would be a reading of ~16.5mV across the emitter resistors (R20+R21).

Output Offset (Spec +/- 1.2mA, +/-25mV)
Before Bias Adjust [No separate Offset adjustment trimpot]
L -1.2 mA / .014 VDC
R -2.7 mA / .007 VDC

After Bias Adjust
L .000 mA / .009 VDC (Can drift up to .019)
R .000 mA / .009 VDC (Drifts down to .005)

(L can drift up to .019. Putting a cooling finger on T1 brings it down. Cooling its partner T2 moves it up. Fun to watch differential gain matching in action.)

A spec for current when measuring the DC offset is unusual; the voltage spec of ±25mV is typical. The ~9mV measurements are acceptable.

 

[Scott75s1974]

These specs are odd, the bias current is often 10-30mA, e.g., a bias of 20mA would be a reading of ~16.5mV across the emitter resistors (R20+R21).

Thanks. I note they were both set significantly higher when I found them. One was more in line with the norm you cite. (I regret not measuring the preadustment bias voltage.) But the SM schematic is accurate. So I presume its specs are accurate as well. I could try returning them to the higher setting (20mA) and see what the voltage does. As it is, there is not much biasing going on at all.

I'll also try to borrow a Fluke to confirm these readings.

 

[Goldie99]

Just spent some time looking at the bias settings Q. as well - you're not the first to run into this issue...

https://vintage-audio-laser.fr/viewtopic.php?t=2995&start=25 (French - but Google translate probably helps).

The top post on the second page, and a few thereafter are the most useful - but in summary - it's suggesting that the measurement with a Triplett 630a meter should be 1.2mA, not 0.2mA. That measurement is not the bias current, but the current shown on the Triplett 630a, under the exact conditions specified, e.g., using the 12mA dc current range only (as the Scott manual is also quite precise on). Converted to a DMM voltage measurement across R20 + R21 (e.g., the more conventional way to measure bias today), a 1.2mA reading on a Triplett 630a would correspond to a voltage of 24mV, and be equivalent to a more 'reasonable' bias current of ca. 29 - 30mA.

In addition, the schematic shows reference voltages of 0 +/- 0.025V above & below R20 / R21 resp. - so up to a max of ca. 50mV across R20 + R21 combined. To assume a 'mid-point' value of say 25mV seems perfectly reasonable - although to be on the safe side, 15 - 20mV would probably be a good starting point.

If you set it too low, e.g., as per the apparent manual specs, then crossover distortion is likely to become an issue, although it should be easy to see if you have a scope.

 

[Scott75s1974]

Just spent some time looking at the bias settings Q. as well - you're not the first to run into this issue...

https://vintage-audio-laser.fr/viewtopic.php?t=2995&start=25 (French - but Google translate probably helps).

The top post on the second page, and a few thereafter are the most useful - but in summary - it's suggesting that the measurement with a Triplett 630a meter should be 1.2mA, not 0.2mA. That measurement is not the bias current, but the current shown on the Triplett 630a, under the exact conditions specified, e.g., using the 12mA dc current range only (as the Scott manual is also quite precise on). Converted to a DMM voltage measurement across R20 + R21 (e.g., the more conventional way to measure bias today), a 1.2mA reading on a Triplett 630a would correspond to a voltage of 24mV, and be equivalent to a more 'reasonable' bias current of ca. 29 - 30mA.

In addition, the schematic shows reference voltages of 0 +/- 0.025V above & below R20 / R21 resp. - so up to a max of ca. 50mV across R20 + R21 combined. To assume a 'mid-point' value of say 25mV seems perfectly reasonable - although to be on the safe side, 15 - 20mV would probably be a good starting point.

Thank you, that is enormously helpful. There is indeed a typo in the SM. The AUDIO TEST pages list .2mA, while the notes on the macro schematic list 1.2mA. Thanks for the translation from Triplett to contemporary DMM. That's close to where it was...and makes a lot more sense in terms of actually biasing those transistors.

If you set it too low, e.g., as per the apparent manual specs, then crossover distortion is likely to become an issue, although it should be easy to see if you have a scope.

I'm completely new to this, but will give that a shot. I'd like to see if its visible. Is a scope watching for the distortion jump the best way to _set_ bias while watching the waveform? Or is it just a check to see what's going on?

I have a very old scope. My Grandfather's Heathkit IO-18, which the internet tells me is rated at 5mhz. Does this seem capable of displaying what I need to see? I have a signal generator to match. I guess the scope will tell me if the signal generator is working before I start.

 

[Leestereo]

...I could try returning them to the higher setting (20mA) and see what the voltage does...

The easiest way to set the bias would be to measure the voltage drop across both emitter resistors (0.82ohms) and adjusting it to obtain the idle current desired (as per Ohm's Law).

...Is a scope watching for the distortion jump the best way to _set_ bias while watching the waveform? Or is it just a check to see what's going on?

I have a very old scope. My Grandfather's Heathkit IO-18, which the internet tells me is rated at 5mhz. Does this seem capable of displaying what I need to see? I have a signal generator to match. I guess the scope will tell me if the signal generator is working before I start.

Using a scope to monitor the waveform is a "nice to have", but not absolutely necessary to set bias and depending on your setup, the crossover distortion may or may not be easily seen.

 

[Scott75s1974]

Once again, thank you for all of the guidance.

Here is the first portion of my Recap list. I still have to do the tone board (hard to inspect) and the radio boards (not many electrolytics).
I'd appreciate guidance that I'm on target. I'll replicate these choices for the rest.
I'm doing Wima films for small values, Muze for Amp boards, Muze BP for signal path and Nich UPW for the Regulator boards. (I just stuck with Nich b/c my head was going to explode with all the choices.) I'm not touching any ceramics.

On a few, I had to jump several voltage levels to find one in the series.
The original installs didn't seem to care much about 47uf vs. 50uf, so I've taken similar liberties here.

My novice criterion for signal path on Amp boards involves carefully tracing the schematic...and generally concluding: if it doesn't connect to ground or power in, it's signal path. I'm using all audio class on the boards just to make sure.

Preamp C2, 102, and 602 are odd. One looks like a big diode, others look like black stacked film but are polarized. I'm assuming they are Tantalums as per this discussion.

Symbol/Per Schematic/ Installed part/ Signal? / Spacing /Rep Type / Mouser #
Preamp
C1/101 100uf 6v 100uf 10v Power 5mm Muze UKZ1H101MHM
C2/102 22uf 10v 22uf .6v Signal 5mm Muze BP UES1E220MEM
C3/103 22uf 25v Signal 5mm Muze BP UES1E220MEM
C5/105 100uf 6v 100uf 10v Signal Muze BP UES1C101MPM
C601 220uf 35v Power 5/10mm Muze UKZ2A221MHM
C602 10 uf 20v Power 13mm Diode shaped Wima MKS4B051004J00KSSD

Driver
C2 2.2uf 22v Signal Wima PET MKS4D042204F00KSSD
C5 47uf 15v 50uf15 Power Muze UKZ1E470MPM
C6, C10 47uf 50 v 47uf 75v Signal Muze BP UES1H470MPM
C11, 12, 14 .22uf Green chiclit Wima PP MKP4F032203G00JSSD

Regulator
C1, C6 47uf 35v 47uf 35v Power Nich UPW UPW1H470MED
C4 22uf 25v 22uf 25v Power Nich UPW UPW1E220MDD
C8 50uf 15v 50uf 15 Power Nich UPW UPW1E470MDD
C11 47uf 16 50uf 15 Power Nich UPW UPW1E470MDD

 

[Scott75s1974]

I have a question about the power to the drivers. They receive full pos/neg rail voltages to the respective PNP/NPN power transistors (pins 1 and 14 in schematic below). But they also receive a separate input from the positive rail (pin 12, left side schematic below), through a very large diode (oriented to allow flow), with a separate 1000uf/55v filter cap (to ground... duplicating the main positive filter cap). There it feeds into the differential pair emitters through a 27k resistor. Schematics have a +.6v reading there.

I don't get what this does. Separate power for the differentials? Some cleaner reference voltage? That diode is very big for just a reference voltage...and its unregulated so no.

The only thing I can guess is that it is a separate power input w/ sep filter cap to isolate the differentials from the voltage drop that the drivers/outputs might put on the rails under a big signal load so the differentials can still do their job independent of the chaos upstream? Like a constitutionally independent judiciary? In that case, capacitor size matters here to maintain _literal_ separation of power ;-). If so it might be worth bumping up that cap a few sizes?

It's currently underside on the chassis. But one other rebuild on line (in German) notes that there is a slot for the 1000uf on the boards that isn't used. He used that slot and put one on each driver board, along with a diode for each (giving 2000uf total). I'm not inclined to do that without a reason. But the traces are there and do exactly what is done in the chassis. I'm guessing the move to the chassis enabled them to bump up the capacity of that capacitor to a size they couldn't fit on the board in 1974 sizes.

 

[Leestereo]

...I'm doing Wima films for small values, Muze for Amp boards, Muze BP for signal path and Nich UPW for the Regulator boards. (I just stuck with Nich b/c my head was going to explode with all the choices.) I'm not touching any ceramics.

On a few, I had to jump several voltage levels to find one in the series.
The original installs didn't seem to care much about 47uf vs. 50uf, so I've taken similar liberties here.

My novice criterion for signal path on Amp boards involves carefully tracing the schematic...and generally concluding: if it doesn't connect to ground or power in, it's signal path. I'm using all audio class on the boards just to make sure.

IMO, this is a good reasonable/pragmatic approach.
Additional considerations: whenever possible, use electrolytic capacitors rated at 25V or higher (the lower voltage rated types have higher ESR and have been reported to be shorter lived). Ceramics are sometimes found in the signal path, e.g., as low pass filter, or high-frequency bypass, in these cases, the preferred types are NP0/C0G. Local decoupling (DC filtering) capacitors can be generally be increased in capacity (as well as voltage); low ESR types are preferred here (better ripple tolerance). Some capacitors that are connected to ground may be in the signal path, e.g., low pass filter capacitor or the DC blocking capacitor in the feedback loop.

 

[Leestereo]

Preamp
C1/101 100uf 6v 100uf 10v Power 5mm Muze UKZ1H101MHM
C2/102 22uf 10v 22uf .6v Signal 5mm Muze BP UES1E220MEM
C3/103 22uf 25v Signal 5mm Muze BP UES1E220MEM
C5/105 100uf 6v 100uf 10v Signal Muze BP UES1C101MPM
C601 220uf 35v Power 5/10mm Muze UKZ2A221MHM
C602 10 uf 20v Power 13mm Diode shaped Wima MKS4B051004J00KSSD

C1/101: emitter resistor (R1/101/) bypass (i.e., signal), replace with 330µF-470µF capacitor to extend function to 20Hz
C2/102: increase voltage, can replace with 1µF film capacitor if desired (input impedance ~420kohm)
C5/105: feedback loop capacitor, BP type preferred, increase voltage rating
C601: increase capacity (e.g., 470µF), low ESR recommended
C602: increase capacity (e.g., 22µF-47µF), low ESR recommended

 

[Leestereo]

Driver
C2 2.2uf 22v Signal Wima PET MKS4D042204F00KSSD
C5 47uf 15v 50uf15 Power Muze UKZ1E470MPM
C6, C10 47uf 50 v 47uf 75v Signal Muze BP UES1H470MPM
C11, 12, 14 .22uf Green chiclit Wima PP MKP4F032203G00JSSD

C1: low pass filter, replace with C0G type if not already NP0 or mica
C2: increase to 3.3µF (to lower F3 to 5Hz, i.e., 2 octaves below 20Hz).
C4: feedback loop, replace with C0G type if not already NP0 or mica
C5: feedback loop, replace with BP type, increase to 100µF
C6/C10: DC filtering, increase to 100µF, low ESR type preferred

 

[Scott75s1974]

C1: low pass filter, replace with C0G type if not already NP0 or mica

Thank you. I really appreciate your generosity in doing this. I'll check what is there. Replacement C0G such as this? FK28C0G2A331J

 

[Leestereo]

Regulator
C1, C6 47uf 35v 47uf 35v Power Nich UPW UPW1H470MED
C4 22uf 25v 22uf 25v Power Nich UPW UPW1E220MDD
C8 50uf 15v 50uf 15 Power Nich UPW UPW1E470MDD
C11 47uf 16 50uf 15 Power Nich UPW UPW1E470MDD

C1/C6: increase capacity (e.g., 100µF-220µF)
C4: increase capacity (e.g., 33µF-47µF)
C8/C11: increase capacity (e.g., 100µF)

 

[Leestereo]

I have a question about the power to the drivers. They receive full pos/neg rail voltages to the respective PNP/NPN power transistors (pins 1 and 14 in schematic below). But they also receive a separate input from the positive rail (pin 12, left side schematic below), through a very large diode (oriented to allow flow), with a separate 1000uf/55v filter cap (to ground... duplicating the main positive filter cap). There it feeds into the differential pair emitters through a 27k resistor. Schematics have a +.6v reading there.

I don't get what this does. Separate power for the differentials? Some cleaner reference voltage? That diode is very big for just a reference voltage...and its unregulated so no.

The only thing I can guess is that it is a separate power input w/ sep filter cap to isolate the differentials from the voltage drop that the drivers/outputs might put on the rails under a big signal load so the differentials can still do their job independent of the chaos upstream? Like a constitutionally independent judiciary? In that case, capacitor size matters here to maintain _literal_ separation of power ;-). If so it might be worth bumping up that cap a few sizes?

It's currently underside on the chassis. But one other rebuild on line (in German) notes that there is a slot for the 1000uf on the boards that isn't used. He used that slot and put one on each driver board, along with a diode for each (giving 2000uf total). I'm not inclined to do that without a reason. But the traces are there and do exactly what is done in the chassis. I'm guessing the move to the chassis enabled them to bump up the capacity of that capacitor to a size they couldn't fit on the board in 1974 sizes.

Your analysis is correct, the additional local decoupling capacitors are to isolate the differential input signal from being modulated by the power supply sag during heavy current demands from the output stage.

 

[Leestereo]

Thank you. I really appreciate your generosity in doing this. I'll check what is there. Replacement C0G such as this? FK28C0G2A331J

Yes, that is an appropriate 330pF C0G replacement; note however, the leads are cut on that one, if you want longer leads you can order these: https://www.mouser.com/ProductDetai...3KoXD5rJ2N9MygfILGh5W3iaZokGRGINDrpxBfQifPg==

 

[Scott75s1974]

C4: feedback loop, replace with C0G type if not already NP0 or mica

Looking at the boards. On the driver C4 and C8 are NP0, the rest are Z5P or U. All the Ceramics on the PreAmp are NP0.
Lower voltage radio boards are non NP0.

Any rhyme or reason seem apparent there? Or just what was in the bin?

 

[Leestereo]

...Any rhyme or reason seem apparent there? Or just what was in the bin?

It looks like NP0 ceramics were used for values ≤47pF; this may have been the upper limit of what was available as NP0 back when this unit was manufactured. Nevertheless, the fact that they use NP0 ceramics at all is indicative of a high quality build, as is the use of glass epoxy PCBs.