Tech21 Character Series [schematic]

[estragon]

Woo Hoo
My British pedal just arrived from Dale Rowley (ebay seller - proaudioland on Ebay) in California USA.
My first impressions on opening the package ( while eating my breakfast) were, that the TECH 21 guys have done a really nice job at packaging this product. You get a sticky label (to help advertise their product), a pressed black tin embossed with TECH 21 in gold writing on it - no cheap brown cardboard box!!!
Upon opening the tin, I was greeted by the pedal, sticky label and instruction manual - the bottom of the tin box even has a small piece of foam plastic to cushion the contents - nice touch TECH 21.
I hastily removed the pedal from it's plastic bag and what is this - a small black plastic quick release battery cover on the bottom of the case - no need to own a screwdriver to change the battery
The graphics are amazing and even the stomp switch and input and output sockets are dull brass coloured to mimic the Marshall gold.
Must get dressed now and get on with the repairs
More later
cheers
bajaman

You must be a kind of monk if you are able to work before even listening to your new pedal!
Congratulations on your new acquisition, by the way.

[bajaman]

First picture of the British pedal PCB
Note - the CH34-5 encapsulated module and the different value boxed capacitors.
Values to follow later
bajaman
members/bajaman/TECH%2021/British/

[estragon]

First picture of the British pedal PCB
Note - the CH34-5 encapsulated module and the different value boxed capacitors.

This module is a real die hard!
Hopefully knowing the internal topology it should be possible to measure the resistors without having to degoop it. I doubt the zeners are of different voltage, as this would not make a significant change. There is no provision for using zeners of different voltage either (for assymmetry), so I guess it is safe to assume the zeners and the opamp are the same.

It would be interesting to measure the voltage that appears on pin 4 of the module when the pedal is powered, just to check if it is different from the "-4" version. It is possible, but perhaps not likely, that the module is the same but with a different P/N just to mislead degoopers. Something to take into account anyway.

Pictures are of very good quality. I can even read the "6G" on the JFETs!

[bajaman]

Well here is what I found with the British
First - the CH34-5 module is identical internally to the CH34-4 module used in the Blonde and the VT Bass models.
I removed the module and thanks to mxrmxr and dimebuGG I was able to measure most of the resistor values (except the one from the junction of the zeners which is open circuit in the CH34-4 module). They are identical to the values measured by mxrmxr. At this stage I was going to put the module back in the mainboard but decided to crack it open just to be sure. I need not have bothered because it is identical internally - the same smd zeners and the same TLC2262 dual op amp.
It looks like Jim was right on here

So what is different with the British when comparing it to the Blonde

First: R4 is 10k and E1 has a 1.8nf smd cap fitted. This gives a much brighter boost in the response characteristic compared to the Blonde which has a smooth 6db boost centred on 1Khz. The British has an additional 8db boost at 6.5khz (2db above the Blonde's 1khz response).

Second: the two box caps ( C5 and C6) are both 10nF (Blonde = 22nF), giving the character control a much higher boost and cut frequency centre of 1600Hz (Blonde = 725Hz). The character control has a +24db boost and -24db cut response at these frequencies (fully clockwise and fully counter clockwise).

Third: R14 is reduced from 47k to 33k in the British - Only a slight difference in the high frequency rollof characteristic ( British = 4.8khz, and Blonde = 4.2khz) I guess this is the difference with the Celestion speaker response model in the British compared to the Jensen speaker response model in the Blonde

Fourth:In the midrange control section C20 and C21 are changed from 10nF in the Blonde to 22nF in the British.
The Blonde boosts and cuts centred on 725Hz and the British much lower on 400Hz.

That's it folks - no other mysteries here no more - next pedal on the block is the Liverpool when it arrives
cheers
bajaman

[lolbou]

So exciting!! Really great!!

[JiM]

So what is different with the British when comparing it to the Blonde

Bajaman, you rock !

It looks like R9 is different : 1k in the Blonde, 3.3k in the British, maybe a jumper in the VTBass.
I wonder if C12 to C16 have different values too ... only R14 seems not enough for emulating a different HP.

Here is a quick summary, feel free to fill the last columns or add anything relevant :

Code: Select all

        | Blonde  | British  | VTBass  | Liverpool | California |
--------+---------+----------+---------+-----------+------------+
ID      |    A    |     B    |    E    |           |            |
module  | CH34-4  |  CH34-5  | CH34-4  |           |            |
--------+---------+----------+---------+-----------+------------+
R4      |  3.3k   |    10k   |    ?    |           |            |
E1      |  open   |   1.8n   |   ~1n ? |           |            |   input &
Cin     |   22n   |    22n   |   10n ? |           |            |  filtering
C2      |   47n   |    47n   |  open   |           |            |
--------+---------+----------+---------+-----------+------------+
C5      |   22n   |    10n   |   10n   |           |            |
C6      |   22n   |    10n   |   10n   |           |            |  Character
R9      |    1k   |   3.3k   |    0 ?  |           |            |    pre-EQ
E2      |  open   |   open   |  open   |           |            |
J1      |    0    |     0    |    0    |           |            |
--------+---------+----------+---------+-----------+------------+
R14     |   47k   |    33k   |    ?    |           |            |
J2      |    0    |     0    |    0    |           |            |
E3      |  open   |   open   |  open   |           |            |  HP sim
E4      |  open   |   open   |  open   |           |            |
E5      |  open   |   open   |  open   |           |            |
--------+---------+----------+---------+-----------+------------+
C20     |   10n   |    22n   |    ?    |           |            |   MID
C21     |   10n   |    22n   |    ?    |           |            |   EQ
--------+---------+----------+---------+-----------+------------+

[mxrmxr]

Well here is what I found with the British
First - the CH34-5 module is identical internally to the CH34-4 module used in the Blonde and the VT Bass models.
I removed the module and thanks to mxrmxr and dimebuGG I was able to measure most of the resistor values (except the one from the junction of the zeners which is open circuit in the CH34-4 module). They are identical to the values measured by mxrmxr. At this stage I was going to put the module back in the mainboard but decided to crack it open just to be sure. I need not have bothered because it is identical internally - the same smd zeners and the same TLC2262 dual op amp.
It looks like Jim was right on here

So what is different with the British when comparing it to the Blonde

First: R4 is 10k and E1 has a 1.8nf smd cap fitted. This gives a much brighter boost in the response characteristic compared to the Blonde which has a smooth 6db boost centred on 1Khz. The British has an additional 8db boost at 6.5khz (2db above the Blonde's 1khz response).

Second: the two box caps ( C5 and C6) are both 10nF (Blonde = 22nF), giving the character control a much higher boost and cut frequency centre of 1600Hz (Blonde = 725Hz). The character control has a +24db boost and -24db cut response at these frequencies (fully clockwise and fully counter clockwise).

Third: R14 is reduced from 47k to 33k in the British - Only a slight difference in the high frequency rollof characteristic ( British = 4.8khz, and Blonde = 4.2khz) I guess this is the difference with the Celestion speaker response model in the British compared to the Jensen speaker response model in the Blonde

Fourth:In the midrange control section C20 and C21 are changed from 10nF in the Blonde to 22nF in the British.
The Blonde boosts and cuts centred on 725Hz and the British much lower on 400Hz.

That's it folks - no other mysteries here no more - next pedal on the block is the Liverpool when it arrives
cheers
bajaman

Fantastic work bajaman Great info as well. I'm surprised you cracked open the goop module How was it for you

The zener in your module, did it have any other markings on it ?

Looking forward to the Liverpool pedal. Should be interesting to see what the differences are in that one

Anyway...... great work. Your a true pro at this

[andregarcia57]

So what is different with the British when comparing it to the Blonde

Bajaman, you rock !

It looks like R9 is different : 1k in the Blonde, 3.3k in the British, maybe a jumper in the VTBass.
I wonder if C12 to C16 have different values too ... only R14 seems not enough for emulating a different HP.

Here is a quick summary, feel free to fill the last columns or add anything relevant :

Code: Select all

        | Blonde  | British  | VTBass  | Liverpool | California |
--------+---------+----------+---------+-----------+------------+
ID      |    A    |     B    |    E    |           |            |
module  | CH34-4  |  CH34-5  | CH34-4  |           |            |
--------+---------+----------+---------+-----------+------------+
R4      |  3.3k   |    10k   |    ?    |           |            |
E1      |  open   |   1.8n   |   ~1n ? |           |            |   input &
Cin     |   22n   |    22n   |   10n ? |           |            |  filtering
C2      |   47n   |    47n   |  open   |           |            |
--------+---------+----------+---------+-----------+------------+
C5      |   22n   |    10n   |   10n   |           |            |
C6      |   22n   |    10n   |   10n   |           |            |  Character
R9      |    1k   |   3.3k   |    0 ?  |           |            |    pre-EQ
E2      |  open   |   open   |  open   |           |            |
J1      |    0    |     0    |    0    |           |            |
--------+---------+----------+---------+-----------+------------+
R14     |   47k   |    33k   |    ?    |           |            |
J2      |    0    |     0    |    0    |           |            |
E3      |  open   |   open   |  open   |           |            |  HP sim
E4      |  open   |   open   |  open   |           |            |
E5      |  open   |   open   |  open   |           |            |
--------+---------+----------+---------+-----------+------------+
C20     |   10n   |    22n   |    ?    |           |            |   MID
C21     |   10n   |    22n   |    ?    |           |            |   EQ
--------+---------+----------+---------+-----------+------------+

thanks!!!!

[estragon]

...Third: R14 is reduced from 47k to 33k in the British - Only a slight difference in the high frequency rollof characteristic ( British = 4.8khz, and Blonde = 4.2khz) I guess this is the difference with the Celestion speaker response model in the British compared to the Jensen speaker response model in the Blonde

That's what I call working fast, Steve.

Just trying to make sure here. Regarding the cabsim, you mean you actually measured C12, C13, C14, C15 and C16 and they turned out to be the same as in the Blonde?

On another subject, note that having R6 different from R9 produces different boost and cut center frequencies, which is something that makes sense in the context of a guitar amp. In the case of the British I'm getting 1.6 KHz boost and 900 Hz cut for the Character control. With the pot centered you get a shelving (tilted) response that has 6dB gain at 10 kHz with respect to 1 kHz and lower which have 0 dB. Quite interesting.

[bajaman]

Nice catch Jim
Yes it makes much more sense having different (lower) cut frequency for the character control - R9 is indeed 3.3k in the British pedal - where would I be without another team member looking over my shoulder
Thanks for that guys.
What a pain in the a#$e it is removing the goop module from the main PCB - not to mention cracking it open.
The smd zener had KV3 and another figure which I could not make out at right angles after it - possibly the maker's date code. I am 99% sure that it is two 3.3v zeners and has negligible effect on the modules performance characteristics. I was able to confirm that there is again (like the Blonde module) no resistance between the common pin on this smd double zener and pin 8 of the module.
Yes - I physically removed all the smd capacitors in the low pass filter sections ( as well as in other parts of the circuit ) - they were all identical in value to the Blonde.
Next stop Liverpool
great team effort here folks - international co-operation at it's best
take care
bajaman

[osiris]

Congratulations and thanks to bajaman, mxrmxr and all the others who have contributed to this awesome thread. It's just like reading a thriller. I really admire your skill and appreciate the effort jou have put into this. There is one question I want to ask, though: Has anyone built and tested the circuit yet?
I breadboarded the CH 34-4 section today and found out that the resistor between Pins 3 and 7 (R106) reduces the gain so much that there is absolutely no distortion, even when the drive pot is at full gain. When you remove R106, everything works like I think it should: adjustable from clean to nice overdrive.
So I believe there must be something wrong in the schematic concerning this resistor and I'm asking you to re-check and either confirm my observation or tell me what I've got wrong.

[bajaman]

I breadboarded the CH 34-4 section today and found out that the resistor between Pins 3 and 7 (R106) reduces the gain so much that there is absolutely no distortion, even when the drive pot is at full gain. When you remove R106, everything works like I think it should: adjustable from clean to nice overdrive.
So I believe there must be something wrong in the schematic concerning this resistor and I'm asking you to re-check and either confirm my observation or tell me what I've got wrong.

I have not physically built one yet (still working on the through hole PCB layout in my limited spare time ), but I have confirmed the value as correct and it is connected as shown on DimebuGG's schematic, and it works fine in the frequency simulation program I am using.
cheers
bajaman

[estragon]

Congratulations and thanks to bajaman, mxrmxr and all the others who have contributed to this awesome thread. It's just like reading a thriller. I really admire your skill and appreciate the effort jou have put into this. There is one question I want to ask, though: Has anyone built and tested the circuit yet?
I breadboarded the CH 34-4 section today and found out that the resistor between Pins 3 and 7 (R106) reduces the gain so much that there is absolutely no distortion, even when the drive pot is at full gain. When you remove R106, everything works like I think it should: adjustable from clean to nice overdrive.
So I believe there must be something wrong in the schematic concerning this resistor and I'm asking you to re-check and either confirm my observation or tell me what I've got wrong.

R106 is OK. Please check you are using the right value (220k) and that C7 (47uF electro) is really tied from pin 3 to GND, otherwise you are introducing negative feedback which effectively reduces gain. As an extra check, you should verify with a voltmeter that the node at pin 3 has a fixed voltage between 5.5V and 6.2V, being 5.8V the expected value, and doesn't vary with the audio.

[osiris]

Thanks, estragon!
You put me on the right track. I had forgotten the 47u cap (C7), which was the reason for the low gain I experienced. Still, I wonder what difference the whole construction around the separate bias voltage for IC2a and the feedback resistor R106 makes. But be that as it may, I'm planning to build a simplified version of the circuit with a FET input buffer, one-OP-Amp speaker sim, Big Muff style tone control and bipolar transistor output buffer. That way I only need four OP-Amps and a reasonable number of components that I can manage with my skills. My build will certainly sound different from the original, but hopefully not worse.
Great thread, great forum, keep it up
osiris

[bajaman]

But be that as it may, I'm planning to build a simplified version of the circuit with a FET input buffer, one-OP-Amp speaker sim, Big Muff style tone control and bipolar transistor output buffer

Looks like the next DIY ready to build project - make sure you take lots of pictures and publish it here on freestompboxes
cheers
bajaman

[Lysander]

Just wanted to chime in and say the research you guys did on these pedals is amazing - i've always liked the Tech21 pedals a lot, and i love getting to know more on how they work.

Can't wait to see the details for the rest of the Character line!

[estragon]

What a pain in the a#$e it is removing the goop module from the main PCB - not to mention cracking it open. The smd zener had KV3 and another figure which I could not make out at right angles after it - possibly the maker's date code. I am 99% sure that it is two 3.3v zeners and has negligible effect on the modules performance characteristics. I was able to confirm that there is again (like the Blonde module) no resistance between the common pin on this smd double zener and pin 8 of the module.

I verified in the sim that this zener arrangement clip around 3.8 to 3.9V away from Vref, so this is about 0.5V less than what the supply voltage would allow, considering the TLC2262 is a rail-to-rail opamp. I tend to think that this improves clipping in some way, otherwise why bother? Anyway I don't have 3.3V zeners right now to check for myself. Also, the sim doesn't show any effect for the zener midpoint resistor, unless you start lowering its value to something below 1k.

Yes - I physically removed all the smd capacitors in the low pass filter sections ( as well as in other parts of the circuit ) - they were all identical in value to the Blonde.

I see. It's just I expected more from the "Voiced specifically for the XXX pedal" statement.

Next stop Liverpool
great team effort here folks - international co-operation at it's best

Amen and amen, brother.

[estragon]

Here I am posting the Blonde and British schems as I redrew them for my personal use.

[bajaman]

For your enjoyment: here are some curves from my simulation - the blue curve is the Blonde and the red curve is the British.
All the controls are set at half way except the character control - the different blue curves show the character control at 10k ohm increments - same for the red curves.
I will post more curves later
cheers
bajaman

members/bajaman/TECH%2021/British/BLOND ... curves.PNG

[sonicvi]

Ok, don't say I never do anything fo y'all. Here's a zip file full of very high res shots of the Bass VT board. There's a lot of redundancy, but better to be thorough and not miss anything. I think everything should be able to be read.
http://homepage.mac.com/sonicvi/BassVTPics.zip