TSL board revisions: what entailed their changes?

Vesperado

Active Member
Joined
May 24, 2011
Messages
331
Reaction score
150
Someone was kind enough to send me a few schematics today regarding the JCM2000 TSL-100 Issue 20 drawings. These are hard to obtain, so I have taken the liberty to share my observations between my amp's boards and the 2007 revisions, aka the infamous "Issue 20".


TL10-62-02
- R 21 : 100R > 33R = Switching time for updated relay scheme?

TL10-60-02
- C 46 : 500V > 1kV = Prevents self-ocsillation.

TL10-61-02
- C 40, mine: "C 39" : 22u > removed = Delays switching relay.
- R 46, mine: "R 1" : 2k2 > 2k2 = "R 1" rededicated and relocated within the circuit as R 46.
- Switching : Opto-couplers > Relays and Diodes = Parts availability?
- VR 4 / 6 : B22K > B20K = Parts availability?
- C 36 / 37 : 63V > 63V = No change.

TL10-66-01
- VR 4 / 6 : B22K > B20K = Parts availability?

Notes:

To improve channel switching via the foot controller I have:

Replaced R 1 with 1k value. Do not relocate. This improves switching time.
Redirect C 39 to ground. Can be clipped. Same.

R 21 I have not touched nor verified. Satisfied with performance as is. Methinks the change in value occured with the change for relay switching over the older opto-coupler design (pre-2007).

Regarding the Bias drift solution:

C 46 was clipped on my board. I installed a 3kV type. Technicians recommend anything stronger than a 2kV type there. [Update: the Issue 20 which I received came with a 6kV type, albeit the drawing has a 1kV. This means that further changes are being implemented on the factory floor.]

C 36 / 37 I have replaced with 630VAC / 250VDC poly film types. This allows the current to flow unimpeded between the drive channels and the mainboard. An aftermarket upgrade is available in the UK which address these ratings albeit Marshall did not.

Tone controls:

These pot values changed from 22k to 20k probably on account of part manufacture. Unless you can get (or make) B25k it is better to leave these values alone.


Concerning TL10-62-02, I have not observed ANY OTHER CHANGES between my Issue 6 and the latest from Marshall. I have been told that some amps have not experienced volume loss/noise running the Effect Loop. My guess is that something took place on the production floor at Marshall which was not recorded in the earlier 2007 revision schematic, similar to what happened with the Issue 5 mainboard having incorrect values. Whether or not these were *special part to order* boards released for purchase under the Part No. MTL10-62-02-01-S is beyond me, the "S" designating "Special". Perhaps someone would like to chime in on this very board who has had experience with the same? IF THIS PART did address these problems then we, amp owners stateside, are at a loss; the distributor cannot source these parts anymore, so I am told.

What does this mean? I doubt it means that an Issue 20 board solved the Effect Loop problem. If anything it means Issue 20 is not the FINAL SOLUTION, but another revision (out of many) which attempted to improve the previous ones. They still fall short. If Marshall has terminated support for this amp, then we must consider it terminated.

Luckily there IS A SOLUTION for the Effect Loop across the board (all Issues #s). I had sent an email to Matt over at Atomiumamps "Europa Mods" to see if he would provide a solution. He did. Six resistors total require changes. Seeing how manufacturer support is gone, I highly recommend contacting Matt for the solution.

One other note, if your amp still has the original power transformer from the manufacturer and you purchased it stateside, it will need replacing. These trannies, as I understand it, were selected on account of being lightweight and able to pass compliance. This saved the manufacturer costs, but left American consumers with the responsibility to replace them. Many amps have experienced failure with these, so if you do not want your amp to fall victim to a costly repair, do replace it. I have a Mercury Magnetics MAR100-P which performs well. Others recommend Classic Tone, too.
 
Last edited:

nkd

Member
Joined
Aug 30, 2017
Messages
75
Reaction score
96
Very useful info, thanks for posting.
 

Vesperado

Active Member
Joined
May 24, 2011
Messages
331
Reaction score
150
Well folks, my Issue 5 board has finally succumbed to the burn-out that we, PCB amp owners, fear. Luckily AES has them in stock for a decent price, so I have one on the way. All that's left for me is to part swap my components over to the new board...
 

Vesperado

Active Member
Joined
May 24, 2011
Messages
331
Reaction score
150
The board arrived a little while ago. I snapped some photos between my Issue 5 and their 20. No hardware came included, no fuses, and NO drawing. There are still carbon comps on the last issue and the soldering of the sockets do not look trustworthy to me. I am kind of hesitant to even touch the new board due to the questionable soldering work of the power valve sockets alone. I plan on rebuilding the board, but I need a board soldered better than this for a rebuild. I think it would be best to contact AES in the morning before moving forward with this project...we'll see what they will do...





 
Last edited:

thetragichero

Well-Known Member
Joined
Jun 6, 2010
Messages
3,428
Reaction score
1,377
Location
upstate ny
board looks like just about every modern pcb amp i've ever opened up, although the flux left on the tube socker solder joints is clear instead of amber colored
get yourself an old toothbrush and some denatured alcohol (sure, everclear or high % isopropyl will work too but i can't really buy either of them by the gallon so i use denatured) and go to town scrubbing. i like to have a rag to wipe after scrubbing
 

Vesperado

Active Member
Joined
May 24, 2011
Messages
331
Reaction score
150
board looks like just about every modern pcb amp i've ever opened up, although the flux left on the tube socker solder joints is clear instead of amber colored
get yourself an old toothbrush and some denatured alcohol (sure, everclear or high % isopropyl will work too but i can't really buy either of them by the gallon so i use denatured) and go to town scrubbing. i like to have a rag to wipe after scrubbing
I could, but what concerns me is the amount of heat applied to the area during production not so much the flux, whoever it was, and putting 525V+ across it may not be such a good idea. I have no idea how they quality check these, perhaps even less strenuous than yesteryear, so I'd rather play it safe than sorry.
 

Vesperado

Active Member
Joined
May 24, 2011
Messages
331
Reaction score
150
AES gave me the option to either replace or return it for a refund. I was planning to send it back and wait for a fresh batch from the manufacturer as I was informed that the whole lot looked like mine. Who knows when the next batch would come?

Convinced that all the manufacturer has done is cranked these out on questionable standards, I went ahead with the project rather than wait to be dissapointed a second time, but taking the necessary precautions to prevent thermal conductivity between pins 4/5 on the output valves. I used a drill, dremel, and diamond file to route out material, being careful not to cut any traces on the solder-side of the PCB.

Unfortunately, when I had prepped the board, I soon had discovered that the assembler went to town with the flux; this made it quite tempermental to clean up the solder underneath the solder joints for the output valve sockets. I tried to soak up the excess solder with wick, which didn't work, so I improvised with some paper and electrical tape to make a smaller hose-fitting onto my solder-sucker. We do not want solder to migrate between the socket-pins! When these boards get hot, and they do, that excessive flux, mixed with solder, will flow to the nearest metal surface it can find...

Cleaning that disaster took almost the full afternoon!

C29 was beefed up to a 400V rating. Issue 5 had 250V.

R18 = 1M on the Issue 20, but = 2M2 on my Issue 5. I replaced it to match my current spec. There are a few other changes on the board as far as components are concerned:

- Switching opto-coupler has been replaced with a Relay incorporating a diode, a polarized cap, R83 replaces R45, and R50/51 have accomodationg values for the relay.

- Two additional connectors (15/16) are not used per a review on a website. No drawing came with the board, unfortunately.

My spec'd main pointers are:

- raise filament BR off the board
- improve DC filament filtering
- beef up higher-current connections with wire and solder throughout
- isolate pins 4/5 V5-8
- Nichicon, nichicon, nichicon!
- Vishay, 2W resistors
- FREDs in HT recto
- R71 > terminals for choke
- change various values here and there to fit my tone

Here is the finished board populated and ready to install:




Once I get the bias set later today, and I am hoping the hard work will have paid off, I am anticipating that the tone will be off a hair as the array of MKP caps will need time to burn in.

Wish me luck!
 
Last edited:

Vesperado

Active Member
Joined
May 24, 2011
Messages
331
Reaction score
150
The rebuilt board works ok. There is less voltage across the filaments. Strange. PT/OT ran cold to the touch after playing for an hour and a half. But I am smelling solder now (and why I stopped playing). I will inspect it once I get another chance in a few days, there might be some flux seeping aroung the pins, still...

...the saga shall continue...

 

Vesperado

Active Member
Joined
May 24, 2011
Messages
331
Reaction score
150
I dug up an old email which contains a digital copy of the Issue 20 drawing, having almost forgotten that someone sent me one long ago...thanks Danny!

I said above that it works "ok" as I noticed a volume drop and some sluggishness, thinking it was the metalized poly caps burning in, well, so I thought late last night. Upon comparison of the drawings I noticed four significant changes, one which I had caught with mine eyes--the value drop at R18, but alas, overlooked R13, R77, and (although I got to it anyway) R68.

[Edit: overlooked because my Issue 5 had a 220k at R13 instead of the published 100k. I've seen values of 100k, 150k, and 220k in this position on Marshall drawings...it's up to the user really...although you get good headroom and signal integrity with 260V (100k) over 175V (220k). R77 is 10k on Issue 5, R68 is 33k. If you want duo-valved operation, not the full quad, 39k works well with the 10k, fwiw. One could modify the bias pot circuit based on part availability should they need to...i.e. PCB mount pots.]

Issue 7 made the changes for the bias circuit, at R68 and R77, 22k and 3k9 respectively, and Issue 20 made the change to signal attenuation, at R13 and R18, 220k and 1M respectively.

Why they decided with Issue 20 to attenuate the SIGNAL and not reduce the GAIN at the valves is beyond me. This is by no means an improvement, nor can it be, and still the manufacturer has left owners with an unresolved Issue since '07...

To top it off, the excess flux has created another problem, as there is need to now suck out the solder from the pins and resolder them with some high-temp type in order to keep the hot mess from dripping all over the place...standards...where did they go?? Every time heat is applied, thermal resistance reduces, and what is the purpose of purchasing a botched board from the manufacturer??

:shrug:
 
Last edited:

Vesperado

Active Member
Joined
May 24, 2011
Messages
331
Reaction score
150
Finally I have the amp up and running; the problem was rectified. I had to suck the solder, clean, solder, suck that solder, clean, solder, suck that solder, solder, then clean. I had to do it an additional turn on about 7 pins, so in all about 3-4 times. I can only hope that the board has not succumbed to thermal degradation, but after applying so much heat to it I cannot reasonably suspect it to last as long had it been assembled with quality solder/no flux from the start. There was a hint of solder migration when I checked, good thing my nose has feeling! I hope Marshall is taking notes here...

I resoldered all major components while the board was back on the bench. Same ordeal.

R77 was changed to 10k metox. Now my DC filament volts are correct. I cannot concede with leaving 3k9 there, as I am convinced the change with Issue 7 was completely uncalled for; it should have been reverted back in 20, imo. **N.B. The purpose for the Metal Oxide 2W is because I have the same type in R68, and aim to maintain a solid bias. So far so good.

I've left R13 alone. Although I may well tinker with it later.

I replaced the new wirewounds (1k x4) with what came on the Issue 5, as they are (a) heavier and (b) have larger surface area for improved heat dissipation; basically a beefier resistor for the obvious reasons. If you look at the rear board in my amp, I replaced the DC recto wirewounds with older 10W types for the same obvious reasons, although I did temporarily mount them on top the chasis, but found that it was not necessary if I simply beefed up the pair and raise them off the obard. Additionally, I drilled vent holes in the steel plate where the dual-triodoi protrude; this helps vent the heat coming off those wirewounds.

The solder used to resolder the sockets was 96 Sn 4 Ag, fwiw.

Bias shows 514V off the anodoi, plate diss. is ~56%, total output power is ~30W if that.

Sounds absolute beast.
 
Last edited:

Vesperado

Active Member
Joined
May 24, 2011
Messages
331
Reaction score
150
With some more down time I am going to make some further adjustments to the Issue 20 board. I can sense a tiddly bit of sluggishness in my pickup response and some low-end looseness to boot.

Here is what I am going to do:

-C9, a plate treble bleed of 100p. With 220k it bleeds everything above 7.2kHz; with 100k, 15.9kHz. I'm going to outright pull it.
-restore R13 to Issue 5 spec of 100k. This should improve the integrity of my signal at the first stage.
-link R26 seeing how they replaced the opto-coupler with a relay, it is no longer needed. Also, the Lead is lower in volume than the Crunch. They overlooked this in their Issue 20. Go figure.
-Try new valves in V2/3.
-adjust Reverb circuit signal strength upon readjustment of Mainboard (R1 Reverb [input], R25/26 Lead [output], and R42 Main [wet]).

V2 is just a dinky coloring to the overall preamp in terms of the overdrives. They reserved it for the Clean channel, following V1B. I pulled C24 so that my Cleans maintain their chime and spank. If one wanted, they could find a suitable value polarized cap and have one leg attach to TR3 and the other to Pin8/R17 with some wire, minding your polarities, and make the Clean tone less boomy. Seems like a worthy effort over the lowsy treble bleed. Remember, positive lobe vs. negative lobe.

Comparing the TSL preamp circuit to the JTM and JCM, it appears that they adhered to the JTM schema, more or less, and abandoned the JCM cold clipper design. The Normal cascades into the Bright with the TSL, not the other way around, for cathode biasing. 1k8 gives a nuetral bias. I opted for a .62 pol. over their 1u, but you can throw this on a switch if you wanted to. Same with installing a cold clipper resistor, put it on a switch. 10k is the JCM800 value.

I am expecting that the removal of C9 will restore picking attack, as there is another global bleed at the end of the circuit before the PI, same corner freq. Ofcourse, I can either restore it or change it provided the results. This may, or may not, require further adjustments to the pre-emph and Volume circuits on both drive channels.

More to follow...
 

Vesperado

Active Member
Joined
May 24, 2011
Messages
331
Reaction score
150
That did the trick...call me skippy!

It was as if I installed hotter ceramics in my axe. Same tubes. Tight, sprite, o the dynamics! Headroom for days--especially the Clean channel! Thunder be cracking out of my cab--wabaamm!

Here is what I did:

Mainboard, Issue 20

R13 > 100k
R26 > linked w/ copper wire
C9 > removed
R42 > 1k (or even linked, all the way up to 2k2 for taste)

Reverb PCB

R1 > 92k (or up to 100k for taste)
C5 > 100u 63V pol. (or to taste)

Lead PCB

R25/26 > 6k8 (the Reverb begins to blend in at about 1.5 on the dial if using a B20k like mine); anything less you might get feedback over the springs.

**N.B. I like my Reverb splashy. The 92k ohm value provides for a large indoor swimming pool--a really nice tone that does well for both overdrive and spanky cleans (provided you still have the original Accutronics unit). You must shield your tank completely and attach a chasis ground to it in order to get clean Reverb. Also, R1 lower than ~82k gets a small shower room type tone, whereas over ~100k a large open venue. 82k gets you concrete tunnel tone, fwiw. The bigger value at C5, the more low-freq. pass-thru.
 
Last edited:

Vesperado

Active Member
Joined
May 24, 2011
Messages
331
Reaction score
150
Now to add in the final touches. Playing the amp for a day has revealed to me some further oddities which have escaped the ear previously. No biggie, I have time ;)

Today I'm going to tinker with pre-emph gain stack couplers, tone stack treble bleeds, and preamp signal couplers. I feel that there is still some room to improve overall tightness and attack.

C27 appears to be a safety feature designed into the circuit with the opto-coupler switching components in mind, should one fail it would not take out the dual-triode at position 3. Again, an arbitrary number of 100n was selected. This allows for all freq. pass-through, which may be the culprit for the always present unwanted/uncontrollable low-end rumble, which do not mistake me, is cool, but you cannot turn it off. I will make this 22n and keep the 630V rating, as this resides in the signal path, and higher ratings provide better resolution borrowing the component from my older Issue 5 board. Plus, its already burned-in.

Looking at the PI circuit, they have a 630V one on the left and a 400V on the right. I will simply move the current C27 part over to occupy C29, taking advantage of that benefit aforementioned above.

On the overdrive boards, there are 4n7 couplers feeding into the treble peakers. I will try 2n2 and see what it does.

I might clip the Lead's Gain Stack Low-pass cap.

I might alter the values on my make-shift Volume Pot Low-pass caps.
 

Vesperado

Active Member
Joined
May 24, 2011
Messages
331
Reaction score
150
Skipping along strong over here :)

Another successful go at improving the circuit on the Issue 20. As I said, the tone was rumbly and boomy, and sure enough I confirmed my suspicions. This is what cleared that up:

C27 100n > 22n
C8 replaced within a tighter tolerance as the one fitted was almost 1n oversized; I canabolized one from the old board.
C29 400V > 630V (refitted from C27)

Touching the Clean channel tone, the loose rumble has been eliminated from the signal. The spank and chime shine through much better because of that.

The Crunch channel likewise, emphasising the character of the natural resonance of the speaker. The Deep is actually useful now, whereas before I never touched it because I couldn't tell with all the constant rumble.

The Lead, same. Did someone back in '98 think that it would result in a warmer tone? It doesn't.

Overall guitar tone is more defined. The Rythm 'bucker is more fat and smooth, the Bridge p'up more dynamic. I can bring up the Neck volume past 8 with both selected and not get loose and rumbly like I did before. So my range of tones has augmented quite nicely because of that. Quite happy about it, actually.

The Gain Stack coupler (4n7) for both drive pre-emph's does change the bass corner. I tried 2n2 which wasn't bad, but I found I prefer (or perhaps hav grown accustomed to) the stock value there. Besides, I also use an ASAT guitar, so best to leave the value alone.

Lead C8 I clipped, which was a 100p that I settled on before, however, I reverted back to it, sort of, but had no 100p left over so I made a 96p and put that in there. With that cap removed, the sound gets VERY woody, to put it in a word, which was quite unpleasant. I think the 100p is THE value for that 2000s tone. 96p seems to add a hint of wood, so the tone control needs to be rolled back some. Perhaps I can fudge a 98p into that place, so as to improve picking attack without the wood, as 100p is great for tone, but at the comrpomise of picking attack with MY axe/V1 valve. My Lead C9 is 120p, another vital aspect of that 2000s tone. It might sound like cutting hairs, but its audibly there alright. Still a very useful tone and good range to dial about.

I adjusted the 82p across the Crunch Volume with a 68p, which yeilds a slightly more classic tone and better picking attack. It clears up some congestion in the highs, too. 82p is the best value I've found for the Lead volume pot, wiper to EE, however.

My Lead V1A coupler is 4n7, like the DSL. This yeilds a woody-esque tonality with the Tone Shifts disengaged, but provide for that thunder with them engaged ala 2000s style. I have just a MLCC in there, as well as in positions 17 and 18 with 740p and 597p, respectively. The tones range from punk, to metal, to rock, to blues, to jazz, to fusion, to, well, basically everything. It is imperative to make these changes as the stock values are way off to get anything decent out of them.

I believe I have now studied this amp enough to understand what it is, or perhaps, aims to be. In essence it appears to be a JTM with a few odd appointments throughout. The Normal is the Amber; the Bright, the Red. They are made to be independent channels which share a Parallel Effects Loop, Spring Reverb, Presence, and Deep control. They both feature a Tone Shift switch which halves the Treble and Mid caps in the Tone Stack. The Amber's Gain Stack is similar to the JTM's Bright channel's, which gives it that notorious "Bright" character, but the Red's has the modified circuit, that is, the JCM2000 "signature". I personally cannot stand it, I think it is awful. To each his own, right?

Rather than dwell on the shortcomings of the JCM2000 line, I've gone through the pains to finish the R&D. 21 years after mine was manufactured, at a point in time in my life where I would be able to accomplish it, to bring it to the repute that it should have arrived at since its inception. I have picked up the pack and ran with it, and now I leave it to the owners out there to pick up the same pack, should they so choose, and it would be a wise choice, whereby this amp can enjoy somewhat of a good repute among those who answer the challenge.

I trust the Father of Loud would approve.
 

Vesperado

Active Member
Joined
May 24, 2011
Messages
331
Reaction score
150
I linked Clean L2 and pulled Clean C1. The difference is notable and doing so keeps your guitar signal pure. The bead/inductor acts as a choke, limiting your bass freqs and the 47p is a low-pass between your guitar and amp, which makes no sense if you aim to amplify your AXE. Now it has not a super drastic effect on guitar tone, yet nevertheless, why color your expensive axe? It doesn't hurt to put a cap in C1 as it kinda helps keep the tone a "modern" one, though 22p would be prefferable: it would improve string definition over the stock value. At any rate, the bead (L2) must needs be removed as it compresses the tone; without it the Deep is twice as effective than before, so a peice of copper wire replaced it for me.

Also, I ditched the A tapers for B linears at VR2/4 on the Clean channel. In tangent with the copper wire above I have never envisioned the amount of GLASS and CHIME and DEEP EXTENSION I have now acheived with the amp's cleans!

A quick note respecting the post above: I managed to get a 97p in place of the 96p, its almost just right...

Also, with the change above: 100k NFB gets good wood, 47k NFB modern djent, and 22k a mean bite. Sizzle sizzle, baby!
 
Last edited:

Vesperado

Active Member
Joined
May 24, 2011
Messages
331
Reaction score
150
I linked R25 today. All three channels now share an equal amount of coloring, or signal degradation if you will: 4 metal film resistors and three MKP caps in the signal path, that's 7 components (compared to the JTM) which alter the signal directly, and pretty significant. At least there is a notable improvement in resolution over the Lead channel now.

Since pulling Clean C1 and linking L2, the treble bleeds called for revision, as 47 pico farad worth of capacitance was restored to the guitar signal. Lead Preamp post coupling cap low-pas filter has augmented to 710p from the previous 597p (100p v20), the Gain Stack bright cap at 120p and bleed at 97p, it's Tone Stack altered to 115p. Crunch Tone Stack bleed now has 115p fitted also. All the woodiness has vanished. I replaced all leaded links with copper wire (signal path only).

The amp is sounding its absolute best now. Had I just cleaned up my input and preamp signal path from the start, I would have saved more time in dialing in the bleed values.

Ok gents, I believe I have sonewhat to boast about now regarding the amp. It may well be the best sounding TSL, third rock from the sun.
 
Last edited:

Latest posts



Top