Marshall Valvestate 8100

[Roly]

I'll soon discover what I can and can't do with amps

One of the great joys of my life is opening things that were never intended to be opened, generally while repairing something that was never intended to be repaired.

I fancy that I can hear, somewhere far across a distant sea, some dude in a suit yelling in pain "You're killing me here!  THROW IT OUT!  BUY A NEW ONE!  MAKE ME RICH!".  And I thumb my nose - I'm on the side of the planet here; repair, reuse, recycle.   :dbtu:

then I played my guitar and measured the voltage across the resistor and measured it referenced to ground

The voltage across the emitter resistor is a proxy for the current passing through it and the transistor.  The voltage to ground is effectively the output voltage, whichever side of the emitter resistor you measure it.

We basically have two situations, static or idle where all the voltages and currents are steady and a multimeter is the most suitable measuring instrument; and dynamic or being driven with signal when the CRO is really a more suitable instrument than a multimeter.


Here are two popular forms of output stage current limiting;

(Note that the output stage consists only of Q2 and the 0.2r emitter resistor, everything else is protection).

Like Enzo, to me that discussion looks like determining at what current the protection would cut in.

To find the ballpark: assuming a 50W amp with +/-35V rails into an 8 ohm load.  Now we know all these numbers are rubbery and subject to a lot of "yeah, but...", however to a first approximation we can estimate the (lossless) peak output current to be;

I = Vrail/Rload

35/8 = 4.375 amps peak.  This is the minimum must-not-operate current because it is within specs.

A typical emitter resistor is 0.22 ohms, so the resulting peak voltage across the emitter resistor will be;

E = Ipk * Re

4.375 * 0.22 = 0.96 volts peak (or not quite a volt peak).

Now as Enzo said, the "given" value for the Base-Emitter cut-in voltage for a silicone transistor is "0.6V" (+/-0.1V in reality).  If we simply connected the protection transistor Base to the emitter resistor it would start to conduct (and clamp the drive into the power transistor) too early, limiting output peaks that are within ratings.  We actually want it to limit for some current above the allowable peak of 4.375 amps, and this might represent a voltage almost double, say 1.0 to 1.2 volts, and this is where the (roughly 2:1) divider across the emitter resistor comes in.

The question being addressed in the other forum is, "at what current does this circuit actually limit?" and the currents quoted look to be just about what I would expect;

It will result in 1.92 volts across R118 , this will give a limit of 1.92 / 0.33 = 5.82 A
For R96 it will result in 2.14 Volts , this will give a limit of 2.14 / 0.33 = 6.48 A.

So the +ve and -ve peak current limits are a bit different, but outside the normal working peak current of 4.375 A, and still well within the capabilities of modern transistors.  What has been prevented is that very high current spike, tens of amps when the output gets shorted under full drive, that kills the output transistors.

So this is worked backwards, starting with the Vbe ("0.6V" nominal) of the protection transistor, multiplied by the resistive divider, giving a trigger voltage across the emitter resistor, and finally the current for that trigger voltage.


The only time there should be any activity in these protection circuits is when the amp is being driven into a load of lower resistance than it was designed to tolerate, e.g. a 4 ohm cab on an 8 ohm minimum.


{Sidebar: on the cct I have the OP transistors are shown with double emitter arrows.  This means that these devices are Darlington transistors;



... which have inbuilt driver transistors, and thus their Vbe is double normal, around "1.2V" nominal.}

[Hawk]

Hey Enzo and Roly,

Wow, thank you Enzo and Roly for your time again. I feel like I should be paying you guys! I went over all your equations and tried them on my own and I understand them. A few more pennies dropping for sure. You can look at this stuff for months and years on your own and not have a clue what it is you're really seeing. So thanks for the virtual apprenticeship :tu:

Thanks also for the info about the Darlington Transistors with the built-in driver transistors, I wouldn't have known that.

The voltage across the emitter resistor is a proxy for the current passing through it and the transistor.  The voltage to ground is effectively the output voltage, whichever side of the emitter resistor you measure it.

When you say a proxy do you mean I can treat as a gauge for  current flow? With multimeter check for high voltage, therefore high current, low voltage low current?



What is the purpose of D4,D5, what is their function in the current limiting circuit? To take only the pos/neg half of the signal? D5--neg. half, D6, pos. half? Both turning on at.7v?


What is the purpose of TR9, R73,R74, TR4, TR7, C49, C47? (I might as well work backwards to fully understand this power amp). Thanks again.

Also, a buddy is bringing over an old tube radio he wants me to fix for him. Can anyone suggest a good forum to ask questions?

[Enzo]

In this very section of the forum, up top are some "sticky" threads, the ones that don't move down over time.  One of those is titled Book ABout Solid State Amps.  Go download it and read it.

By measuring the voltage across any resistor, you can determine the current flowing through it.  That is Ohm's Law, one of the most fundamental concepts in electronics.

Tr9 and the rest of your list are the bias circuit.  They provide what I call a voltage space between the bases of the opposing output transistors.

[Roly]

I feel like I should be paying you guys!

Insert Card Here...

When you say a proxy do you mean I can treat as a gauge for  current flow? With multimeter check for high voltage, therefore high current, low voltage low current?

Exactly.  Current, voltage and resistance are locked together via Ohms Law.

Your older moving coil or d'Arsonval type meter actually responds to current, so voltage must be implied by the large resistance in series with the meter.

Modern digital meters typically only measure voltage, so current must be implied from a small resistance in parallel with the meter.

Similarly with resistance measurement a moving coil meter puts a fixed voltage across the terminals and measures the resulting current, while a DMM applies a constant current to the terminals and measures the resulting voltage.

"E equals I multiplied by R" said Georg Simon Ohm - and you can bet your life on it.

What is the purpose of D4,D5, what is their function in the current limiting circuit? To take only the pos/neg half of the signal? D5--neg. half, D6, pos. half? Both turning on at.7v?

Ohdearohdearohdear.  On the (fuzzy) circuit I'm looking at <8080-pwr.gif> D4 and D5 are drawn the wrong way around!  Both are supposed to be pointing downwards, not upwards - as drawn the protection won't work.

They are there to protect the associated clamp transistor from being reversed biased on the "other" half cycle.

What is the purpose of TR9, R73,R74, TR4, TR7, C49, C47?

See;
http://www.ssguitar.com/index.php?topic=3592.msg28780#msg28780

Also, a buddy is bringing over an old tube radio he wants me to fix for him. Can anyone suggest a good forum to ask questions?

Do not twiddle any of the many presets - these require special equipment to set.

Basic superheterodyne radio receiver and waveforms.


Most electronics faults are common, simple, and fairly easy to fix, however there are some considerable differences between an audio amplifier and a radio receiver (I assume AM only), and a bunch of different concepts and techniques that are required.  If the problem turns out to be simple, e.g. a valve not lighting up, or in the audio section, then you're away, but ahead of the audio section there are a number of other stages;typically an Intermediate Frequency amplifier with tuned LC loads working at 455kHz, before that a Radio Frequency mixer stage, a Local Oscillator running around a megahertz, and perhaps a RF preamp stage (as above).

Radios made for the consumer market have a long history and are not helped by the commercial competition which saw very many designs with all sorts of odd circuit arrangements with the aim of getting maximum performance with minimum cost - and some of the worst cost cutting would make your eyes water.  There were also Patent wars where manufacturers did things differently because they had to, or pay.


By all means have a look (run a mile if it's an AC/DC set, they are killers), check that the chassis is grounded for your safety, then just the usual stuff, valves lighting up?  HT voltage?  Any signs of life in the audio?  Can we see signals anywhere (for which a CRO is very handy indeed)?

Same old story, signal flows from antenna connection to speaker, how far is it getting, and why not?

But having said that, don't be surprised if you very suddenly find yourself well out of your depth.

My first job in electronics was production line servicing AM radios, so I have repaired literally thousands.  Most service techs get to be expert in one or two particular fields, get to know radios, TV's, VCR's, cellphones, laptops, or microwaves and other kitchen appliances, &c&c, and often not a lot outside their area.  I have covered a lot of fields; design, production, industrial, bio-med, off-grid power, theater/film, AM/FM/UHF-TV broadcast, &c&c, and tried to be an "electronics generalist" so the list of things I've worked on (and normally fixed) is huge.

I'm quite happy to hand-hold you through a radio repair, but you need to clearly understand that this opens a whole new bag of concepts (and frankly, I don't think is possible via forum or email if the problem is in the IF or RF stages).

Now we warn folks that building electronics projects can be highly additive, but we don't often mention that if you want to be any good at electronics you had better be prepared for a lifetime of reading, study, and a bit of research (even if you are only ever going to do guitar amps).

You want a flying start? - then NEETS, here;
http://www.fcctests.com/neets/neets.htm

A good grounding in Module 1 is vital, but after that you can just browse Modules as the need arises ('tho you may well find that Module 17, Radio-Frequency Communications Principles, first needs an understanding of Module 12, Modulation Principles, and perhaps others).


Yagotta stay curious 'bout stuff.

[Enzo]

I recently bought a couple old nice looking tube table radios for $5.  I'll either fix them or turn them into cute guitar amps.  But if I make them work, frankly, what I expect to find is bad caps.  I doubt the IF and RF stages will be out of tune, at least unless someone was in there messing with them.   If it is a tube radio, it is at least 40 years old, so very likely the electrolytics are dried out.  Depending upon age, chances are also good the coupling caps are leaky.  ANy of the old wax covered ones I can just about guarantee are history.  Disc ceramics are probably OK.  Film caps may or not be, one determines that easiest by checking for DC on the downhill side of them.

For many years I restored old tube jukebox amplifiers, and that was the majority of the work I did on them, replacing most of the caps.

Here is a clean looking copy of the Marshall schematics:

[Roly]

Thank you Enzo, that is so much easier on my poor old eyes.

{And here we note that D4 and D5 are drawn the right way around - but now they have forgotten the values for R97 and R117!  (sigh)   "No circuit is without errors".}

When it comes to repairing anything I have observed that commonsense and a good set of eyeballs can take you far - i.e. paying attention helps a lot.

[J M Fahey]

but now they have forgotten the values for R97 and R117!  (sigh)   "No circuit is without errors".}

You sure? 

What about:



hint: the same board is used in 2 different amps, the 8080 combo (80W/8 ohms)  and the 8100 head (100W/4 ohms) .

Not only they need different rail voltages, but also different short protection.

[Roly]

And my paying more attention would have helped.   

Okay, bum rap, that was off the screen in the x2 magnification I was looking at it.  Still, they did manage to get the diodes the right way around this time.

[Hawk]

Roly, I kept trying to insert my card into that image but no luck! I'll keep trying! haha!

Roly, thanks for posting that radio info and the schematic and the information you've gleaned from working on radios.  Also, the current limiting info and link are great as well as the Neets site. Great info indeed!!

Enzo, I have downloaded that book you suggested and have read about 70 pages, and there are sections I have to keep re-reading but it is a great book and I will finish and re-read for sure.

So Beyond the Output Transistors we have resistors and caps between the OT's and the speaker jacks, plus some caps. Are these covered in the solid state amp book? Or can you give me a basic understanding of their importance? Thanks.

Also, What does O/P stand for? I believe it's the middle rail separating the +VE/-VE.

How long has this forum been in operation? I'm tempted to print off all this info but rather than do that and save some trees I'd rather return to it electronically a month/year/decade down the road. Is there a way to bookmark this post as it will eventually descend further down the post line and be tougher to locate. Can it be saved electronically? Thanks.

[Enzo]

Where do you see "O/P"?

In one context it may mean output, while in the context of a forum thread it may mean original poster.

[Hawk]

I'm referring to the O/P at the bottom of Roly's current limiting diagrams at top of page. Thanks

[Roly]

I'm referring to the O/P at the bottom of Roly's current limiting diagrams at top of page. Thanks

In this context "O/P" = output, the amplifier "half-rail".


Working backwards.

How long has this forum been in operation? I'm tempted to print off all this info but rather than do that and save some trees I'd rather return to it electronically a month/year/decade down the road. Is there a way to bookmark this post as it will eventually descend further down the post line and be tougher to locate. Can it be saved electronically? Thanks.

If there is one thing that forums generally lack it's indexing, being able to find what you want in ways other than Google.  Proper indexing of topics has to have a human input from somebody who knows a bit about the craft, sundry wrinkles of terminology that are, or are not, conceptually linked.

The mechanics are there to be used.
Each forum, tread and post has a heading, and that heading contains the URL of the item (right click copy link location, paste, e.g. your post)
http://www.ssguitar.com/index.php?topic=3707.msg28813#msg28813

The page number contains the thread page;
http://www.ssguitar.com/index.php?topic=3707.0

The heading bar;

Solid State Guitar Amp Forum | DIY Guitar Amplifiers >
    Solid State Amplifiers >
    Amplifier Discussion >
    Marshall Valvestate 8100

Also contains URL links.

As you can see above I make free use of hotlinking content.  To me that is the whole idea of the internet.

In Cyberspace the "distance" is the retreval time.

Every single post on every forum is simply an available item, and all you have to do is point to it - from anywhere.

Any page on the net may consist almost entirely of links to external content.  In fact I ran for a short time an edited index to a forum.  I didn't want to stop people going to the forum, but a group of people wanted to present an edited version ordered by topics.

You can make up HTML pages for your personal use, or topic sub-folders in your browser bookmarks.

So Beyond the Output Transistors we have resistors and caps between the OT's and the speaker jacks, plus some caps. Are these covered in the solid state amp book? Or can you give me a basic understanding of their importance? Thanks.

Okay, mid-right we have the power transistors driving the half-rail via a couple of very low value "emitter" resistors.

Next we have a low value resistor in series with a cap to ground.  This is the Zobel stability network.  It is to prevent parasitic HF or VHF oscillation modes in the output pair.

Then we have a small inductor and resistor in parallel.  This is to isolate the output stage from any capacitive loads (such as crossovers or piezo tweeters) that may cause instability.  The resistor is to damp the self-resonance of the inductor (and is often also the form for winding the coil).

According to Doug Self the values of these components is not critical.

[g1]

  Roly, perhaps you could comment on the origin of the term "half rail".
To someone who has only ever seen split supply designs, it may not make sense without context.   

[Roly]

Thanks g1, I was hoping it was implicit, but you're right, it should be nailed down.


In the amp circuit above the output or "half-rail" is the bit between the speaker output terminal (22uH inductor) and the 22k Negative FeedBack (NFB) resistor.


We also have two supply "rails", +38V and -38V with respect to ground.  ("rail" simply means that it is a distributor (of power or signal) and has quite a few connections; "buss" is sometimes also used in a similar context, audio mixers, power distribution).

In between we have the output "rail" connected to the output pair emitter resistors, the speaker, and the Negative FeedBack (22k above).  This output rail is also known as the "half-rail" because when the amp is working correctly it adopts a voltage half way between the supply and ground, or half way between the split supplies (+/-100mV).

This is actually a very important test because (apart from not lighting up your speakers - which is why we keep hammering "disconnect the speaker") if the half rail is right it is a strong indication that the whole output stage is working correctly - it has to be to achieve this "rebalanced" condition (the OP stage is basically a giant power op-amp that rebalances itself to the midpoint because of a lot of DC NFB - if something in this NFB loop isn't working properly then the amp is highly unlikely to "centre" or rebalance the half-rail).

On first encounter the half-rail voltage can also give you a strong indication as to where to look first.  If, say, the upper OP transistor is shorted then the half-rail will be close to the +ve supply, similarly for the lower transistor, and if it is at some other odd voltage it suggests that both OP transistor have blown open.  Like a pulse the half-rail voltage can tell you a lot about the patient with only one measurement.

In the early days of transistor amps it was fairly common to use a single supply;



...where the output would bias to half the supply voltage, hence "half (supply) rail", and needs to be DC blocked by a large output capacitor.  You still find this arrangement in smaller combo amps, particularly older ones of Asian origin.  (The output cap is more commonly between the OP stage and the speaker, with one side of the speaker grounded.)




The trend a bit later was to use a split supply;



...which eliminated the need for a specific output DC blocking cap because the "half rail" was now half way between +ve and -ve, i.e. at ground.  This is by far the most common arrangement these days, and particularly in higher powered amps.


Actually there isn't a lot to choose between the two arrangements, but Doug Self does make an interesting observation that in the latter the speaker is DC coupled to the output stage, and if the OP stage should fail you can get speaker combustion.

Some people (generally Hi-Fi buffs) don't like having a big electrolytic in series with the speaker with all those audio amps passing through (because, ya know, capacitors - they tend to have a thang about capacitors) , but you can redraw both arrangements to show that for audio frequencies it's a distinction without a difference.  In the split rail design the output currents also flow through the power supply filter caps anyway, it just isn't as obvious.

Ground
If you examine the circuit fragments above you will notice that the ground is in different places.  Which bit of the circuit is actually grounded to the chassis is really a bit arbitrary.  You may be using a metal speaker jack that has one side connected to the metal case.  With the single supply it then makes sense to call the -ve end of the supply "ground".

It is not unknown to come across 1970's Asian PNP single supply designs where the supply +ve is grounded (and the negative supply coded red - caution!).

If you are using a split rail supply design then the more logical place is the mid-point of the filter caps, but it is also possible to ground the half-rail output and let the supplies flap up and down (as one manufacturer does - to me this is the tail wagging the dog, but I'm sure they have their reasons).

[Roly]

I just grabbed this circuit at random to illustrate a couple of points about the output section;



... but on closer examination it contains a couple of real design bloopers. 

One error means this circuit would not work at all, and even if corrected the other error would mean it would then not work at all well.

No prizes for the techs, but can you spot them Hawk?


{proving once again that you can't just take any circuit on the internet at face value.}