Scott R75s rebuild questions

Thanks for the responses to these questions.

A question to improve my understanding of how these circuits work.
The value of -7mV at pin 13 is simply the DC offset of that amp + half of the 22mV drop you set for the bias, so -18mV + 11mV = -7mV
The measurement of -7mv at the collector of the PNP power transistor/pin 13 ... I was thinking this meant that that transistor wasn't getting biased. But I realize I'm not thinking of that being tied to the power resistors R20 and R21. I'm just imagining it as an independent output. So since it's tied through R20/21 to the collector of the NPN power transistor, it's being pulled negative. So am I correct to conclude that only the sum of that bias voltage matters, not either "half" measured to ground? Thus, I should not conclude that a negative voltage reading read against ground at pin 13 means the power transistor is not being biased?

(I think my faulty imagination is to think of things in more TTL like signals which is what I played with as a kid. I'm trying to reimagine them as trusses from my statics class. Never had a circuits class alas.)
 
Last edited:
Is the ".20 mV" a typo, i.e., should be 20mV or 0.02V? Assuming that is the case, the 10 minute delay after power turn-on is not unusual, in fact waiting 15-30minutes for voltages/temperatures to fully stabilize is recommended.

That's helpful. SM said 5 minutes. After about 30 minutes the bias measurements increased to 30mV. I played some music at fairly high volume and checked everything again and readjusted for .02v. When things are fully heated up, the questionable board can be biased properly at about 75% of turn. At 5 min, it had to be pegged to its extreme and barely hit .02V. So I guess I can relax about this. (But perhaps I should do exactly what the SM says? Adjust at 5 min. And let it rise to .03V later. Since surely they knew about what would happen afterwards?)


AFAIK, a very low ESR for C6 or C10 wouldn't affect the bias; maybe you are referring to C6 or C10 "leaking" DCV, which would increase the measured DC offset. I think the same effect would occur if the protection circuit transistors TR5/TR6 had leaky junctions.
Sorry for my fumbling with new terminology. I mean if those old caps began to pass current like resistors--so "leak."

Agreed, a problem with D3 would impact the bias, but since the problem followed the driver board and not D3, it doesn't appear that D3 is faulty.
With the driver boards back in original locations, the weird board is now reading -.76v on the -Diode pin 6 instead of .000V. Other board same reading in either slot. I'm not sure this matters.

What was the actual bias reading? If the +ve and -ve probe locations were switched with each other does the bias read as positive.
Actual full bias is -.020V when measured according to SM: pin 13 pos, pin 15 neg. Switching probe reverses polarity. Black on chassis ground yields: -.011 on pin 13, and -.031 on 15.

TR3 is the only transistor that got hot on the boards. Did so on both, but hotter on weird board. It's meant to take some heat. Has a built in sink. Like a junior version of a TO-220 case.
Also the collector voltage point on TR3 which I neglected to measure. Spec: -.6v Good board: -.54v Weird board: .000v

I'd like to figure what is pulling the various points on this board negative. If it seems caps are likely suspects, then I'll go ahead and place the order and start to replace. If it seems likely it's some transistor. I'll order those as well. None are particularly expensive. And aside from differentials, not much hassle to replace.

But if it seems like in the range of acceptable to you all, I'll let it go...and get on with the recapping. Thanks again.
 
Last edited:
Maybe its a mistake to think of those overly neg voltages as the cause. Perhaps it's just that for some reason TR4 (not the power transistor it drives since it follows the driver board when switched) is not particularly responsive to biasing. (In itself, or due to something connected to it.) So that would require jacking the bias current up, which would overbias a healthy TR7 and its power transistor and pull the PNP side bias negative with it? So a weakness on the positive side, not an excess on the negative.

If so, I could check that by adjusting the bias just for the NPN side, measuring that side to ground. If it matches that half of the bias from the good board at a similar trimpot setting, that would suggest that it is a problem with TR4. Does that make sense?

BUT, from my voltage measurements, there is not a huge difference in the actual voltage on what I take to be the actual bias feed in points: the bases of TR4 (.016 lower relative to good board) and TR7 (.030 higher). But those differences do correspond to the negative bias on the bias test points.
 
Last edited:
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?
Did you ever get to the bottom of this problem?
 
Did you ever get to the bottom of this problem?
I assumed the startup thump was simply life w/o a protection relay.
This post -- Rush or surge at turn ON - whrump in speakers --suggests it can be an imbalance between rails at startup, which might go away when I replace filter caps if they have aged differently.

Or not, depending on how big the additional draw is off of the positive rail for various other power draws around the chassis.

This might suggest *against* increasing DC filter caps on various boards. As designed, it has about 1300uf more in DC filtering capacitors on various sub-boards that draw from the positive rail than the negative which only feeds the drivers. I had planned to about double this. The tradeoff might be thump at startup vs. functioning under heavy load.

I'll be able to test this as I do the recap. Once the new PS filter caps are in, I'll start it up w/o the filters for the other boards. Then with the boards installed. That should give a pretty clear answer.
 
Last edited:
Thus, I should not conclude that a negative voltage reading read against ground at pin 13 means the power transistor is not being biased ?

Not in my opinion. Biased & operating 'normally', under DC conditions, I'd just be looking for a ca. 0.6V drop between the base voltage and the emitter voltage on TR4 - the values you noted on your schematic were 0.540V (0.555 - 0.015), and 0.546V (0.539 - -0.007). They both seem fine to me - also don't forget, the schematic voltages are indicative values, measured with an analogue voltmeter not a DMM. They're not spec. values (except maybe the DC-offset & bias numbers).

When things are fully heated up, the questionable board can be biased properly at about 75% of turn. At 5 min, it had to be pegged to its extreme and barely hit .02V. So I guess I can relax about this. (But perhaps I should do exactly what the SM says? Adjust at 5 min. And let it rise to .03V later. Since surely they knew about what would happen afterwards?)

I don't believe the 5 minutes is critical, especially not while you're (presumably) working on it with the covers off, and have much more airflow to the amp / heatsinks... Personally, I usually monitor bias settings for at least an hour, sometimes more, just so I'm sure it's reached a stable thermal equilibrium. It makes little sense setting until it's at equilibrium - not in this case anyway.

Actual full bias is -.020V when measured according to SM: pin 13 pos, pin 15 neg. Switching probe reverses polarity. Black on chassis ground yields: -.011 on pin 13, and -.031 on 15

I don't follow that. Actual bias is -20mV or +20mV when measured according to SM ? If you measure -11mV at pin 13, and -31mV at pin 15 (both vs. GND), then pin 13 should be +20mV versus pin 15.

TR3 is the only transistor that got hot on the boards. Did so on both, but hotter on weird board. It's meant to take some heat. Has a built in sink. Like a junior version of a TO-220 case. Also the collector voltage point on TR3 which I neglected to measure. Spec: -.6v Good board: -.54v Weird board: .000v
But if it seems like in the range of acceptable to you all, I'll let it go...and get on with the recapping

Not sure why TR3 would be so hot (?), or why a couple of voltages seem to be zero some times...

You did say something above about boards in either 'slot' - if they're a push fit connection, did you already deoxit all the board & slot contacts ?

Post #103, I'm afraid you lost me.
 
BUT, from my voltage measurements, there is not a huge difference in the actual voltage on what I take to be the actual bias feed in points: the bases of TR4 (.016 lower relative to good board) and TR7 (.030 higher)

Just re-reading your post #103 - I'm not sure you're understanding the bias current correctly ?

As far as I can see, in a very 'simplistic' but hopefully illustrative, view of the world..... TR4 & TR7 are the 2 drivers for the 2 outputs TR1 & TR2. Now, starting (say) with TR4, there's a small base current, into the base of TR4, which is controlling how much current is passing from TR4(C) to TR4(E). The higher the current passing TR4(C) to TR4(E), the higher the base current being 'pulled' from TR1(B), that in turn determines the current passing from TR1(E) to TR1(C).

The TR4(E) current, and the TR1(C) current, both combine at pin #13 (bias test point 1). That's the bias current being measured via the (milli)voltage drop across R20 + R21.

If you follow the current further, effectively everything goes straight through R20 / R21, on to bias test point 2 (pin #15), and on to TR7(E) & TR2(C), ignoring any feedback via R6. At the TR7 driver, there's a small current out of the base, that's controlling the current passing from TR7(E) to TR7(C), that in turn pushes current into the base of TR2, and determines the current being passed from TR2(C) to TR2(E).

The base currents into TR4 & out of TR7, are essentially balanced, so the bias current effectively passes from the positive supply rail at pin #1, via TR4 & TR1, R20 & R21, and TR7 & TR2, to the negative supply rail, at pin #14.

Anyway - maybe not the most elegant of explanations, but hopefully it helps & doesn't confuse.... current flow(s) on the other side of TR4 & TR7 can be considered analogously, to see how the bias current is being controlled / varied by RV1, D3, and temperature of the outputs. Remember Kirchoff rules...

Time to 'duck' I think.....
 
Just re-reading your post #103 - I'm not sure you're understanding the bias current correctly ?

As far as I can see, in a very 'simplistic' but hopefully illustrative, view of the world..... TR4 & TR7 are the 2 drivers for the 2 outputs TR1 & TR2. Now, starting (say) with TR4, there's a small base current, into the base of TR4, which is controlling how much current is passing from TR4(C) to TR4(E). The higher the current passing TR4(C) to TR4(E), the higher the base current being 'pulled' from TR1(B), that in turn determines the current passing from TR1(E) to TR1(C).

The TR4(E) current, and the TR1(C) current, both combine at pin #13 (bias test point 1). That's the bias current being measured via the (milli)voltage drop across R20 + R21.

If you follow the current further, effectively everything goes straight through R20 / R21, on to bias test point 2 (pin #15), and on to TR7(E) & TR2(C), ignoring any feedback via R6. At the TR7 driver, there's a small current out of the base, that's controlling the current passing from TR7(E) to TR7(C), that in turn pushes current into the base of TR2, and determines the current being passed from TR2(C) to TR2(E).

The base currents into TR4 & out of TR7, are essentially balanced, so the bias current effectively passes from the positive supply rail at pin #1, via TR4 & TR1, R20 & R21, and TR7 & TR2, to the negative supply rail, at pin #14.

Anyway - maybe not the most elegant of explanations, but hopefully it helps & doesn't confuse.... current flow(s) on the other side of TR4 & TR7 can be considered analogously, to see how the bias current is being controlled / varied by RV1, D3, and temperature of the outputs. Remember Kirchoff rules...

Time to 'duck' I think.....

Thanks for this. I appreciate your time. It's helpful. I get the flows through the transistors. My imagination starts to falter when I deal with what's happening when both sides are tied together across R20 and R21. I don't want to take too much space for this, but perhaps I'm confusing a simple meaning of bias --putting a small current through a transistor-- with the bias setting of the AB amplifier which is setting the handoff between the PNP and NPN push/pull functions of the circuit.

I'll take your word for it that this circuit is ok, and move on with the recapping. Again, thanks.
 
When you set the bias & dc-offsets, the amp is operating on dc only, there should be no ac signal present. You're setting the steady state dc conditions in the amp, so it's in the 'optimum' state, ready to amplify an ac signal.
 
Did you ever get to the bottom of this problem?
Ok, I've finally begun the recap.

I replaced the main power filter caps. Original were 4700uf Elnas. Which measured at 4830 and 4690. Still right on the nose for capacity. 35mm case.
I replaced with Nichicon Muse Goldtone 6800uf. Which measured in at 6230 and 6220 respectively.
(Recall from above that these same components can handle 10000uf in the R77 version. So this was a moderate upgrade.)
These were 30mm, so used foam tape to line clamps.
Thanks to whoever's post I read about soldering an eye connector to the snap-in lead. That was necessary for the 5 wires feeding off of the positive rail.

Immediate improvement! (So something was clearly wrong with them even though they read dead on for capacitance.) Thump/waterfall at power on is 75% gone. With lower noise at startup I can now hear two smaller waterfall curves on the left channel. So probably bad caps on that driver board.

I'm awaiting the big order for the caps for the various circuit boards. Thanks again for all the guidance.

IMG_0556.JPG
 
Last edited:
I'm working through the recap, board by board. So far so good.

As I get to the driver boards, I'll recap and recheck the bias to see if that fixes the problem.
The bias trimpots are 275 ohm, in series with a 47 ohm resistor. Replacements are not easily available. Things in that range are seldom stocked.
Closest I've found is a 200ohm sealed multi turn (measures at 194ohm). I could pair this by replacing the 47 with a 130 ohm. That would give me a similar total series resistance. But the adjustment range would be narrower. I'd have to check and see if this works in real life.

But is there any obvious a priori reason I should not do this?

Schematic available in message #52.
 
Last edited:
I've completed the recap except for the pre-amp. (The voltages I needed were out of stock at Mouser…and well, I also forgot to order for that board.)

There is an enormous improvement. All thump and startup/shutdown noise is now gone. The amp seems to have a very low noise level. Volume at full on null input has audible hiss only near the speakers. I can clearly hear the background recording level hiss on some very low volume level recordings--such as extremely low dynamic sections of Arvo Pärt--Litany.

There is still a significant, but largely silent speaker movement at startup and shutdown. (Unbalanced rails, since positive feeds 5 different boards most with their own local DC filter cap, negative rail only feeds the drivers.) I’ve read this is ok. Is it? I suppose I should just power up with no speakers selected.

The only remaining sound is a very brief hum at startup, which I’m guessing is unfiltered DC coming through in the fraction of second until filter caps charge.

I’ll post pics of the various recapped boards w/ final parts list.

Regulator Board

Sorry, the heat sink prevents a close up of the recap.

C1/C6 Increased from 47uf to 220uf, Nichicon UPM, UPM1H221MPD
C4 Increased from 22uF to 47uF, Nichicon UPM, UPM1H470MED1TA
C8/C11 Increased from 50uF to 100uF, Nichicon UPM, UPM1H101MPD6TD
C9 Largish black ceramic, .22uf 12 v. Replaced with Wima, MKP4F032203G00JSSD

These smallish electrolytics were by far the most in need of replacement. They were consistently 2-3x times spec. The large, low voltage black ceramic measured dead on, but I replaced because I wasn't sure what a 12v ceramic package is.

IMG_0598.JPGIMG_0618.JPG
 
Last edited:
Driver Boards:

I followed @Leestereo ’s suggestions for upgrades here:

C1 330pF, upgraded to NP0. RDE5C2A331J0M1H03A *See note below
C2 2.2µF tantalum box, upgraded to 3.3 µF Wima film, MKS4D043304J00KSSD (PET, PP was too big to fit)
C3 1000µF, Nichicon UHW1J102MHD *See below.
D1 UF4004

C5 47µF, Signal, upgraded to 100µF bipolar, UES1V101MPM1TD
C6, C10 47µF, Power, increased to 100 µF, UPM1H101MPD6TD
C11, 12, 14 .22 µF green chiclit film caps. In general, I left the films in place. But the .22µF ones frequently had coatings with a gap on the top. Like the dip wasn’t thick enough. I worried these might cause problems. I replaced with Wima film MKP4F032203G00JSSD. The spacing on the Wima’s was long and the leads are short, but I was able to bend a knee underneath to decrease the pin spacing and it all worked out nicely. Sorry no pics.

*Upgrading C3 to a 3.3 µF film was a very tight fit. It only worked because I also upgraded its neighbor C1. That ceramic was nearly microscopic and its fine wire enabled me to bend its leads horizontal and emerge around the bigger C3 which is covering its board territory. It's the little blue dot next to the large Wima red box in center of the second photo.

IMG_0609.JPG IMG_0608.JPG

*I also moved the DC filtering back onto these boards. There is a separate filtered positive input to provide independent power to the differentials.

Parts are marked C3 and D1 on the board, but they are not on the SM schematic. These boards were modified from initial build at the factory--unsoldered, still had flux marks vs. the rest which was fountain soldered. The jumper that replaced them _is_ on the SM driver board and schematic page. Those parts were shown on the macro-schematic showing links between boards. Diode and larger 1000µF filter cap were shared underneath on the chassis by both driver boards. Fed in through a separate pin. I replaced with one 1000µF on each board. The old cap was measuring at 1500µF+. So not much of an increase.

IMG_0619.JPG

Here are some shots of the parts that were removed and their connections:

IMG_0599.JPG IMG_0600.JPG IMG_0601.JPG
 

Attachments

  • image.png
    image.png
    16.8 KB · Views: 44
Last edited:
Tone Board:

C# Original, Replacement
C3/102 2.2µf (Tant), Nich Muse Bipolar ES1H2R2MDM (perhaps I should have done film--Wima PET MKS2B042201F00KF00 would fit)
C4/104 22µf, Nich Muse Bipolar UES1H220MPM
C5/105 220µf, Nich Muse Bipolar UES1V221MHM1TO
C7/107 22µf, Nich Muse Bipolar UES1H220MPM
C8/108 1µf (tant), Wima PET Film MKS2C041001F00JSSD
C601 220µf, Nich UPM Low ESL UPM1H221MPD
(I kept local DC filtering caps to replacement level to avoid increasing startup imbalance on rails. Since the imbalance is in the design, I'm not sure this was necessary.)

@Leestereo also recommended the following replacements which I did not do.
This board is not removable/wire-wrapped to wire harness. I was a bit paranoid about breaking one of those wires, so I decided to minimize replacements and leave the original films and ceramics. The wires seemed fine actually. (I was overly paranoid from some brittle wires in a Heath transistor checker I tried to use. A new one would snap with each repair of the previous snapped one.)

C1/101: 0.033µF (upgrade to polypropylene type, e.g., ECW, MKP, PHE426)
C2/102: 0.0027µf (upgrade to polypropylene type, e.g., ECW, MKP, PHE426)
C9/109: 330pF (replace with C0G type if not already NP0 or mica)
C10/110: 180pF (replace with C0G type if not already NP0 or mica)
C11/111: 0.1µF (upgrade to polypropylene type, e.g., ECW, MKP, PHE426) Wima PP Film MKP2D031001F00JSSD

IMG_0597.jpg
 
Last edited:
A detail on the Driver boards I've not posted, but that's important:

Pin 8 in the center of the boards is soldered. So that has to be addressed if you are going to remove them. I used a desolder iron to get most of the solder out. Then I eased the board off of the barbs on the plastic pin supports, reheated the pin, and gently pulled until the boards came off. Then wen't back and cleaned up the pins with the iron.
 
Last edited:
It really does have amazing quality. So many labor hours must be in it. Even the chassis must have taken hours. Heavy gauge aluminum. All holes drilled, many squares and rectangles stamped out. There are many of these b/c of the daughterboard pin connectors. All deburred. Some countersunk. Final uniform belt sand on blank, then brake-bent into shape. It's a bit gratuitously nice. That final uniform planishing is just not necessary. The heatsinks likewise--drilled, deburred and ground before anodizing. In an age before CNC mills, this must really have employed a factory full of labor. The point to point wiring isn't always beautiful, but other than that it's really hard to find a cut corner.
I agree. The workmanship on the Scott Professional Line is top notch--better than the 350R, for example.
 
A question about inputs on this receiver. There are no aux inputs on a receiver of this age. The manual lists the tape inputs sensitivity as 35mv. I'm generally playing music via the headphone outputs of an iPod or a MacBook. Both, are I believe 1v line level outputs. Obviously this works. I don't top them out b/c I assume that would be inviting distortion in the on board amps from the sources.

But my question is: what is the best level to put into the tape input from these digital sources? I assume the volume buttons reproduce an audio taper. Should I keep the iPhone/Mac outputs low... aiming for something like 35mv? (Or am I misunderstanding sensitivity?)
 
Last edited:
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

Mouser now has the caps I need for preamp in stock. So I'm placing the final order and returning to this rebuild. I have a few last questions.

On C2/102 you recommend switching to a film, which I plan to do. These caps are in series with R9 tied above the board. The total hole width is about 16mm. Since it's the preamp, I was trying to go with PP for the caps. But the Wima PP's are 22mm, the Wima PETs are 5mm. I might be able to fit the PPs in but it's a tight board and I've read large film caps can sometimes cause (inductive?) interference. Any thoughts on a suitably sized/quality film I can fit in this space?

IMG_1026.jpg
 
Last edited:
I have also been thinking of replacing the driver bias trimpot with a multi turn. (They are working now, but are very touchy.)

It's an odd size: 275ohms. The track of the trimpot is in series with R9 a 47ohm resistor.
After much searching, I've found a 250ohm Bourns multiturn at Arrow.
My plan is to replace the trimpot and increase R9 to 75ohms. This will bring the total series resistance in the same.

The only difference I see is that I'll lose 25ohm of adjustment. From the current setting of the single-turn pots, I don't think this will be a problem.

Does anyone see any reason this won't work?

driver.jpg
 
Last edited:
Back
Top Bottom