Welcome to Solid State Guitar Amp Forum | DIY Guitar Amplifiers. Please login or sign up.

March 28, 2024, 10:42:52 AM

Login with username, password and session length

Recent Posts

 

Teisco Checkmate 21 Solid State amp schem + help

Started by galaxiex, July 27, 2014, 10:06:34 PM

Previous topic - Next topic

Roly

Quote from: galaxiexAhhh Ha Ha Ha ...hahaaa gaspchokesnortcoughwheeze....

Sorry... I just spewed hot coffee out my nose and all over the keyboard....

If it hadn't been for Intel and Microsoft we would have had keyboards with detectors and airbags to prevent this sort of thing by now.   ::)


Yep yep, Doug Self, Pat Hawker, Baxandall, the whole Newnes/Wireless World crew.


Quote from: galaxiexthis amp has never had a hum issue

This typically has to do with a shared earth path between a supply rail and the path around the tranny, rectifier, main filter cap.  The "right" place to make supply and return connections to the PSU is at the main filter cap terminals, that way the resistance in common between the charging and discharging paths is limited to the ESR of the filter cap.


Quote from: galaxiex4. All 4 pot grounds now go direct to board near their respective circuitry.

+1   :dbtu:


Quote from: galaxiex5. Input jack is *not yet* insulated but I ran a jack ground wire all the way back to the main filter cap anyway.

With an insulated jack you take its earth return to the AC common ground point for the first stage, not back to the central ground point - it then goes there on its specific ground, but first reference the signal to the amplifier ground, not via a long loop back to the single-ground point (it should still be quiet, but...).


Quote from: galaxiexInstalled both bypass caps C14 and C15 (not sure how their values affect gain, if at all) used 1uF for C14 and 10uF for C15.

These control the local AC NFB around the pre-driver and driver stages.  As their values are reduced the low end rolloff comes up in frequency, bass cut.  I'd be inclined not to dick with them for the moment because your f/b sounds like it is mid-hi freq "", n'est pas?  So we need to go looking for a high frequency feedback path we don't want, and apart from common grounds, that can mean unwanted capacitive coupling.  The grounding around the pots doesn't sound too flash, and localising these controls and their grounds would be something I would have been looking at next.  {the old layout isn't up to the new gain, some re-engineering is required, but this continues to be a great learning experience when you start off with a sow's ear and some prospects.  :lmao: }




nope, that cct won't fly.

Now need to mod the input to have a "Half Rail" Vsup/2 provider.  1:1 voltage divider across the local supply, bypass mid point to AC ground, this is now "pseudo-ground" for the op-amps, and they are effectively running from +/-supplies half the whole supply.

This allows simplification of the input biasing, just a single 2M2 to this new half-rail point, and it provides for low-Z gain arrangements.

I've cascaded the gain and buffer stages to avoid the need for an output mixer.






PS
cross posts
good new, yeah the 330 sets the loop AC gain

I wouldn't be content with that 'tho, tamed, but not yet unconditionally stable.

If you say theory and practice don't agree you haven't applied enough theory.

galaxiex

#91
Quote from: Roly on September 01, 2014, 01:51:21 PM
I wouldn't be content with that 'tho, tamed, but not yet unconditionally stable.

Yeah, seems like lotsa tweaking still to do...
If I knew how to use/apply some of the math I've been reading, I'm sure it would go much easier/faster.  ;)

Anyways....

Keeping up to speed here...

1. Input jacks now insulated, and grounded to Vol pot ground. Which goes to board near Q1, not ideal but works.
1a. Existing Input buffer presently grounded on input jack, so, ok.

2. Bypass caps C14, C15 now 10uF and 47uF.
With the 100uF at C15 tooo much bass response, maybe if I played a Bass Guitar. :lmao:
(sows ear heh heh heh  :))

3. New schem of input buffer(s).


--------------------------------------------
Random thought(s), feel free to ignore....

Ever has there been such effort and re-engineering expended on such a (started out as) piece of crap amplifier...
Only a fool or insane dummy would do such...   :duh  :lmao:

The end result (if we ever get there) Is going to be soooooo coooool!  8|

More random thought...

How about a "Fuzz" circuit for J3?
Re-purpose those germanium transistors I pulled off the board. Yeah! Built in Fuzz Face!  :lmao:

Could label the front panel...

1. Normal
2. High
3. Fuzz

Sorry... I can't help it... the Muse is with me and I'm just inspired...

besides... i have this huge (95x47cm) stripboard for mounting the input buffer(s) and reverb brick, (when we get there)
and all this room near the input jacks...

Fun Fun Fun...  :)

Edit; does that R5 1K go to ground, or the half rail?
I drew it to the half rail but thinking it goes to ground...
Oh wait, I figured it out.. corrected schematic...
This look good?

Oops, C7 is backwards.

Is there a convention for an input gain stage like this?
Buffer first then gain? or vise versa? Does it matter?

Yet another question...
Can we safely increase the gain of U1b? As configured its X10, can we go higher?

I have a guitar that I really like, it plays great but has very low output. I don't want to change the p-up's
This is why I keep wanting to add a High gain input to this amp.

For the record, the guitar is a 60's Teisco/Silvertone and I want to keep it "stock"
I'm a small guy and these guitars have a short scale and are very light, easy for me to play.
I don't use it all the time, simply because of the low output.
I know I can stick a pedal in front and get gain that way, but I don't (like to) use pedals much.
My main guitar is one of those Made in Mexico Nashville Tele's, great guitar.
If it ain't broke I'll fix it until it is.

Roly

Quote from: galaxiexdoes that R5 1K go to ground, or the half rail?

Half rail.


Quote from: galaxiexIs there a convention for an input gain stage like this?
Buffer first then gain? or vise versa? Does it matter?


In2 --> x10 --> In1 --> buf -->
Always got the buffer, option of x10


In1 --> buf --> In2 --> x10 -->
Always got the x10, option of the buffer


The second doesn't make sense.


Quote from: galaxiexCan we safely increase the gain of U1b?

"Safely" in the sense of it catching fire or exploding?  Yes.

"Safely" in the sense of it oscillating like crazy?  Well, mostly.

The Maximum Available Gain is of the order of 100,000 times.  You can probably go up to x100 before it oscillates like crazy (and you should be able to play with one hand only).  Actually attempts at ultra gain like this generally discover that the guitar pickups are slightly microphonic and you start to get "pickup squeal" very like PA mike feedback.  You can't do any harm, so try a few different values and see how far you can push the envelope.


If you follow the x10 amp with a 1k-10k resistor then a pair of silicon diodes (e.g. silicon power or signal, even LED's) in anti-parallel* to ground you should get a pretty effective fuzz.  (*in parallel but opposite directions to each other)
If you say theory and practice don't agree you haven't applied enough theory.

galaxiex

#93
Quote from: Roly on September 02, 2014, 01:27:50 PM
Half rail.

Got it, I'll fix the schem.

Quote from: Roly on September 02, 2014, 01:27:50 PM
The second doesn't make sense.

Right, I meant in general for other designs...
I haven't looked at enough SS amp schematics to have a grasp on what the "normal" design conventions are for clean and gain channels.

Quote from: Roly on September 02, 2014, 01:27:50 PM
You can probably go up to x100 before it oscillates like crazy.....
.....discover that the guitar pickups are slightly microphonic .......

.......try a few different values and see how far you can push the envelope.

Yep, I've had that from pickups that are not potted, and even some that are....

Ok, I'm guessing R8 10K is the one I increase to up the gain?

Fuzz;
Right, 1 to 10k + diodes to ground, seen that lots.  :)

Another rambling random thought...

Some of this stuff (electronics) is slowly coming back to me.  ;)
Big difference in just copying something out of a book or wherever,
compared to actually understanding what the circuit does and being able to apply it to other stuff.

Edit; I'm getting this wired up right now (sans fuzz circuit) and I noticed something....

Do we still need R7 100K from pin 7 of U1b?

We originally used the 100K on the output of the buffer to get the trem to work.
R7 is not needed for that here.

Now, I can see that with nothing plugged in at all,
it provides a ground path for J1 suppressing any noise from an otherwise open jack.

Correct?

(see, I'm lookin at stuff and trying to understand how it works  :))
If it ain't broke I'll fix it until it is.

Roly

Quote from: galaxiexI haven't looked at enough SS amp schematics to have a grasp on what the "normal" design conventions are for clean and gain channels.

There are all sorts of variations, but one way channel switching is done is like this;

In --> preamp stage -->Ch sw--> -------------------->Ch sw --> Master
                            --> gain+diode clipper-->

A footswitch drives electronic signal switching which inserts extra gain and a diode clipper in the middle of the preamp.  This is essentially a built-in fuzz box with an electronic (transistor/FET/relay) double pole double throw switch at each end - just a complicated version of stomp switching.


Quote from: galaxiexDo we still need R7 100K from pin 7 of U1b?

Consider: we have a guitar plugged in to J1 when we switch on.  Without R7 there is no way for C5 to charge up, so its "groundy" end is at half rail with the output of IC1b.

Now you unplug your guitar and the amp input is now suddenly connected to a voltage of V/2 while C5 charges.  "SPLAT!!!".

With R7 C5 charges, so when you pull out your plug the input will switch to "ground volts".  Bit of a click maybe.  This is why you see those "useless" 10 Meg resistors on every audio switch contact to ground - it's to keep all the caps charged and mission-ready for when you change a switch, and they are all charged up to their working voltages ready to go, no ugly splats.

Quote from: galaxiexCorrect?

Correct.   :dbtu:

Note a little feature of your arrangement is that the pre-amp buffered guitar signal is available on J3 as a recording (dry) output.




A few op-amp basics...

The voltage gain of IC1b is the ratio of R5 to R8 (a.k.a. Rshunt and Rfeedback, Rs, Rf)

Voltage gain Av ~= R8/R5 (use for large ratios, high gains)

Av == (R5+R8)/R8 (accurate, use for gains below ~20)

N.B. that this is the same as the voltage divider formula, upside down.

Vout = Vin * R2/(R1+R2)


e.g. "x100"

Rf = 100k, Rs = 1k

(100+1)/1 = 101 - close enuf for audio, only 1% error.

but say now "x3"

intuitive from above, Rf = 3k3, Rs = 1k

(3.3+1)/1 = 4.3 - hummm not even close.

(2.2+1)/1 = 3.2 - better

(1+1+1)/1 = 3 - exact with 2x 1k in series for Rf.


{In a Negative FeedBack path everything is upside down, so an attenuation becomes a gain in the forward path.}

By transposing;

Av == (R5+R8)/R8

... you can find gain, R5 or R8, given the other two.
If you say theory and practice don't agree you haven't applied enough theory.

galaxiex

Thanks again for all your help Roly. Very much appreciated. Just wanted to say that publicly.  :dbtu:

Got the buffer/gain board done. Works a treat!  :)
X10 gain seems to be plenty for now. (maybe too much, but the rest isn't "quite right" yet...)

Got some other things done too, will update later, for now some pictures....

Excess wire lengths will be trimmed later, got some new pots on order.
If it ain't broke I'll fix it until it is.

Roly

Just went through all the pics.  Man that looks a whole lot better (well the bits you built, anyway).

Mains is safer, PCB add-ins are great.   :dbtu:

One thing, those control leads, no wonder you got instability with all that hair.  Short and Direct is the rule, or screen/shield them.


"Frank-In-The-Box"
If you say theory and practice don't agree you haven't applied enough theory.

galaxiex

#97
"Frank-In-The-Box"  :lmao:

I deliberately left the wires long cuz I knew I'd be making changes... I guess that's not really necessary.  :loco
Can change stuff even if wires are the "correct" length and if too short... wires are (relatively) cheap and can be replaced.

I'll correct the lead dress tonight.  :)

BTW that white wire from the buffer to amp is the signal. If I touch it I get a radio station...  :o
I'll definitely shield that one.

I have six colors of wires I use like this...

Red - power + OR -
Black - ground
Green - tone
Blue - trem
White - signal
Yellow - used for speaker out and misc. signal

Output transistor wiring.

Black - base
Red - emitter
Green - collector
If it ain't broke I'll fix it until it is.

galaxiex

Better?  ;)  :)

The grey wire from buffer to amp is now shielded.
At full volume not playing I can still faintly hear a radio station....

Was thinking when it's all said and done I will strip the chassis and blast it.
(I have a blast cabinet) and a nice paint or even powder coat?
The original anodizing is stained and discolored on much of it.
If I could find a place local that does anodizing, that might look nicer than paint.
(if not too expensive)
If it ain't broke I'll fix it until it is.

galaxiex

#99
Here's the other changes I mentioned a couple posts above...

I followed the values in the data sheet (see pdf attached) for the 2N2148 output transistors.
The data sheet shows with 16.5 Vcc a maximum signal power output of 15 watts.
We have @ 25V on the rail.

But I'm not sure what we have for input impedance of each base.
They show 65 ohms.
Is that the impedance of the 2 transformer coils going to each base?
I'm sure that affects output power.....

Is the above for a split supply where each rail is 16.5V ?

Cuz on the right side at the bottom they show...
"Typical operation in single ended push pull amplifier circuit"
"DC collector supply voltage 16.5V"
Does that mean single supply?

I'll go check the voltages at the output....

I was initially afraid to try those R values as they are much lower than what was there originally.
Decided what the heck and... they work.
Transistors get barely warm. So little that it's hard to tell if they are warming up at all.
Everything else stays cool as well. Transformer too.

Amp is slightly louder and cleaner with those R values installed. (8 ohm speaker for now, when I get the cabinet built, it will be 4 ohm)

Edit; oops forgot to add the gain/buffer to the schem...  :-[
If it ain't broke I'll fix it until it is.

Roly

Quote from: galaxiexBTW that white wire from the buffer to amp is the signal. If I touch it I get a radio station...  :o
I'll definitely shield that one.

Not surprising considering there is a diode right at the input - you're half way to a Crystal Set.   :lmao:

Screened lead, ferrite beads, and 100pF cap to ground, all anti-RF.


Quote from: galaxiexanodizing

I think you will find that coating is "cadmium-passivated" on steel, perhaps better to get it re-electroplated.

Quote from: galaxiexBut I'm not sure what we have for input impedance of each base.
They show 65 ohms.
Is that the impedance of the 2 transformer coils going to each base?

You can't just measure the resistance or impedance of the base of a transistor.  It's effectively a diode to the emitter, and as such has a complex behaviour, acts like a non-linear resistance with applied voltage.  Here it is the current into the base junction that is interesting and (slightly) more tractable.  There is the hope that the upper and lower cells will tend to cancel out any lumpy impedance questions in the driver transformer.  These two drive secondaries are normally wound bifila, meaning wound on together as two twinned strands so they have the best possible matching.




Quote from: galaxiexI was initially afraid to try those R values as they are much lower than what was there originally.
Decided what the heck and... they work.

We are putting a lot more current through the bias divider chain.  This means that the base currents are a much smaller proportion of the whole, so bias voltage regulation is better.

With the lower resistor reduced to 3.9 ohms this reduces the drive source impedance into the base because this resistor is effectively in series with the base drive winding.

"Just warm" is exactly what germanium semiconductors should be when working.  You can just tell it's on.

The voltage across R29 and R30 would be interesting because that will tell us what the idle current through the output transistors is.

Dunno about 4 ohms.  Let's give this a close scrute working out on 8 ohms first.  Just because you now seem to have a good DC condition, there are AC conditions that aren't nearly so friendly, things called SOAR curves, Safe Operating ARea of duration of high current and voltage at the same time into reactive (read real world speakers) loads.


{I would love to have some accurate AC signal voltage measurements at the onset of clip around that driver transformer.  By knowing these we can work out it's turns and impedance ratios, get an idea of what sort of load the driver is working in to on the primary (and so we explore backwards).}


Dead computer PSUs from your local computer shop, all the project cases and hookup wire (in colours) you will ever need.  HV caps and diodes too.   :dbtu:
If you say theory and practice don't agree you haven't applied enough theory.

galaxiex

#101
Quote from: Roly on September 05, 2014, 12:29:45 PM
....you're half way to a Crystal Set.

:lmao:  :lmao:  :lmao:

Yes, I know diodes are great at rectifying RF signals...  ;)

Amplifier stability;
So... a few changes made with stability in mind.
In order of changes made...

1. Put back original R and C values around Q1. Amp sounds good, barely audible hum (I think was always there but so slight I didn't worry about it)

2. Added C26 for RF suppression. Seems good, no more radio station.

3. R24 and C16 (near T2 on schem) are actually physically and electrically close to Q1, Q3, Q4 and T2.

4. Added R41 and C12 down by the trem circuit. This is probably the most dramatic change, so-far as stability goes. Trem still works and is quieter too. No more "thump" "thump" "thump" as it pulses.

But now have hum issue!  :grr

With all controls at 0, no guitar plugged in, amp hums medium loudish...
Start turning Vol up and hum gets quiet-er and at Vol 6 is almost gone, then increases from 6 to 10.
Quite loud at 10.

So I made a mistake when installing C6 1000uf at the buffer supply rail.
Accidentally connected C6 to -25V rail and the junction of R4 and R11 instead of ground.
(I forgot to install C6 at first... :-[ Now don't remember if hum was there before or after installing it wrong)  :-[  :-[  :-[ Rookie mistake!
Turned amp on and was greeted with a VERY loud hum.
Immediately switched off. Possible I damaged U1?
I only have the one TL072 right now, so can't try a new one to see if that's the hum problem.

The hum is quite nasty but if you can ignore it, the amp is actually quite loud and clean sounding (under the hum noise).

Also:
The gain channel may need reducing to X5 gain, as the present X10 is a bit much.

New schem with some voltages added.

Edit;
Hum is the same with or without guitar plugged in.
If it ain't broke I'll fix it until it is.

Roly

I think C25 is supposed to be between the colector and base of Q1.


Quote from: galaxiexWith all controls at 0, no guitar plugged in, amp hums medium loudish...
Start turning Vol up and hum gets quiet-er and at Vol 6 is almost gone, then increases from 6 to 10.
Quite loud at 10.

The volume pot is now sharing a ground return pathway with a fairly high hum current.  In this case it's less likely to be an unintentional ground path so much as excessive current through it due to a fault, say a boofed op-amp.  Finding how hum currents are getting into your signal pathway can sometimes require a bit of detective work.  First get it working again.  Break the signal path to see where it's coming from.
If you say theory and practice don't agree you haven't applied enough theory.

galaxiex

Quote from: Roly on September 08, 2014, 03:44:18 AM
I think C25 is supposed to be between the colector and base of Q1.

Yep, put it on the board correct, drew it wrong. (smashing head against wall emoticon)

Was going to work on this tonight, but got a box in the mail with the MOD Kits Trill Tremolo diy pedal in it.  :)  8)

So built that instead.  :tu:
If it ain't broke I'll fix it until it is.

Roly

If you say theory and practice don't agree you haven't applied enough theory.