KT 88 project request

[KevinOConnor]

Hi Guys

Ian's post explains the difference very well.

If you ever played through a Marshall plexi or 800, the EQ doesn't do much. On the other hand, playing through any typical Fender amp like a Twin Reverb from the '70s has a much more effective EQ. Marshall's is driven by a CF; Fender's is plate-driven. The EQs in both circuits is pretty much the same schematically and were you to try swapping the specific Fender and Marshall values, the CF drive would still be ineffective, so the problem is the CF NOT any difference in tone stack values.

TUT6 (2008) explains why the CF is in there in the first place, which you can blame Fender for, and why it has persisted to this day, which you can blame Ken Bran for. It also explains as Ian did above why it should not be used in MI for this function. There are much better places for a CF than to drive the EQ.

Have fun

[pdf64]

...If you ever played through a Marshall plexi or 800, the EQ doesn't do much. On the other hand, playing through any typical Fender amp like a Twin Reverb from the '70s has a much more effective EQ. Marshall's is driven by a CF; Fender's is plate-driven. The EQs in both circuits is pretty much the same schematically and were you to try swapping the specific Fender and Marshall values, the CF drive would still be ineffective, so the problem is the CF NOT any difference in tone stack values...

I'm not sure about the above?
My take is that the Marshall arrangement seems less effective mainly due to the 33k slope resistor, and the mid control being wired as a pot; changing the bass control to 500k 10% also seems to help to improve the user interface, in a guitar amp application.

I think that a tinker with Duncans TSC may be seen to bear this out; using the Fender model, the range of bass control is about the same, and treble control range looks to be greater, if the source impedance is lowered from the CC 38k to a CF 1k3.

Another way of looking at the less effective Marshall tone control arrangement is that (together with the presence and Celestions) it prevents the user from screwing up too much the characteristic Marshall sound / it's intentionally limited to facilitating only 'fine tuning'.
As the flipside of having a big range of tone control is that it can make getting the 'right sound' tricky.

[KevinOConnor]

Hi Guys

pdf64: You may not be sure of the above but I am and have been for decades. I made the changes and trials in hardware and players and other techs agree with the facts above since they could hear the difference. As I said, it makes no difference about the actual tone stack values - the CF-drive kills dynamics and effectiveness of the EQ.

There was no trying to restrict the EQ performance on Marshall's part, rather they simply copied one form of the many forms of the Fender Bassman. That amp was used to test a lot of different circuit ideas and a few of these had feedback around the mix stage. It is well known that driving a feedback loop using a low-impedance will make the loop more accurate as the buffering takes the loading of the loop itself off the gain element. That circuit was not popular so the loop disappeared but the wiring for the CF remained, since it was simple and the tube was there anyway. So, as TUT6 states, the CF is a remnant of Fender's experimenting and is otherwise nonbeneficial to the tone.

Players always prefer the plate-driven EQ performance over stock when a Marshall (or any other amp) is rewired this way. Hundreds of players preferred this since their EQ finally did something. It is a funny thing with some though, that if you tell them what the change is then they doubt their own impression since there might be some "original mojo" lost - this despite the fact they were unhappy with the stock sound and requested a mod.

Hans and Franz would say "hear me know, believe me later", but it's easier to believe now...

Have fun

[pdf64]

...pdf64: You may not be sure of the above but I am and have been for decades. I made the changes and trials in hardware and players and other techs agree with the facts above since they could hear the difference. As I said, it makes no difference about the actual tone stack values - the CF-drive kills dynamics and effectiveness of the EQ...

Hmm, not sure about that either, opinion to the contrary (ie concerning the tonal contribution and benefit of a DCCF) is documented, eg The Valve Wizard

I don't see what the mechanism may be by which dynamics and eq efficacy is reduced (by it being preceded by a CF)?

...There was no trying to restrict the EQ performance on Marshall's part, rather they simply copied one form of the many forms of the Fender Bassman...

Note that the early Marshall tone stack values are directly lifted from the published 5F6A, see http://bmamps.com/Schematics/fender/..._Schematic.pdf and http://bmamps.com/Schematics/marshal...5_lead_45w.pdf
Whereas later Marshalls eg http://bmamps.com/Schematics/marshal...d_50w_100w.pdf changed a few values, eg slope resistor to 33k, which a tinker with TSC indicates tends to reduce the range of control.
Hence it seems reasonable to conclude that Marshall's changes to the 5F6A tone stack values may have been intended to make it easier for the user to get a good sound / reduce the likelihood of the amp sounding bad due to a poor choice of control settings.
My finding is that the 5F6A arrangement can be made to sound unpleasant, eg awful blocking distortion at the EL34 control grids with high treble and presence settings, whereas the same settings with the later arrangement avoid that.

Whatever, it's a fun debate! Thanks for your thoughts so far.

[nickb]

Here are some numbers comparing the range of the treble and bass controls for the two stack variants and with 1K and 33K source impedances. I chose what I considered to be perceptually important high and low frequencies of 125Hz and 2Khz at which to make the measurements.

Marshall Stack dB

	Bass Range	Treble Range
Low Z	10.3	7.4
Hi Z	11.6	7.2

Fender Stack dB

	Bass Range	Treble Range
Low Z	12.2	17.9
High Z	12.4	17.7

From this you would conclude that there is not much difference due to the source impedance. The Fender stack does have a much bigger range of control on the treble, as Pete said, quite possibly giving the sort of effects he mentioned.

[catalin gramada]

Hi Nickb. Can you imagine how impedance vary in those two situation to realize how hard those triode run in plate and cathode follower mode.please? could be important to know if this aspect dynamical affect.that considering signal a plied aspects as atack.decay and so on...

[nickb]

Hi Nickb. Can you imagine how impedance vary in those two situation to realize how hard those triode run in plate and cathode follower mode.please? could be important to know if this aspect dynamical affect.that considering all signal a plied aspects as atack.decay and so on...

Very roughly, if you take the value of the slope resistor you won't be too far off for an estimate of the impedance that each stack type presents to the driver i.e 33K for Marshall and 100K for Fender.

This might help to explain why the cathode follower was chosen by Marshall.

[KevinOConnor]

Hi Guys

pd64: The blocking distortion et al that you refer to in a Fender amp is to be expected. There is no voicing accommodation for being overdriven since Fenders were designed to be used clean. Marshall had to make the circuit brighter by rolling off bass and emphasising treble to make the distorted sound nonflatulent. Player feedback drove these modifications.

Have fun

[Gingertube]

I think the point of my post above was missed.
The Duncan TSC just tells you what the tone stack does in isolation from the driving stage - yes you can give it different source impedances to see what happens but that still doesn't explain the drive stage/ tone stack interaction for an anode drive stack.

Say you want a mid cut function. You set the tone stack pots to give that, the tone stack impedance to 0V will be low at those mid frequencies (shunting more of the mids to 0V).
That lower impedance is in parallel with the anode load resistor which will lower the gain at those mid frequencies.
That is, the driving stage will produce less signal at those frequencies to begin with and then that reduced signal will be cut again by the tone stack.
OR
Say you want a treble boost, the tone stack impedance to 0V will then be higher at those treble frequencies (less treble shunted to 0V).
The anode load resistor in parallel with that impedance will be higher and the driving stage gain will be higher at those frequencies, producing more of those frequencies and then less of them will be shunted to 0V.
So the tone stack and its driving stage together will have a larger range of control.

This effect is dependent upon the impedances. With the very low output impedance from a cathode follower you do not get this tone stack + driving stage interaction. The tone stack is "isolated" from its driving stage.

Another useful (but not completely accurate) way of thinking about the anode driven stage + tone stack is that the source impedance into the tone stack is not constant but varies with frequency according to where you set the tone stack pots. The Duncan TSC doesn't allow you to model that.

Cheers,
Ian

P.S. Whenever you use a simulator it is important to understand what it is NOT going to tell you as is to understand what it is going to tell you.
This can often be worked out by examining the basic assumptions the simulator works with.
For the Duncan Tone Stack Simulator it assumes 2 things:
A flat frequency response driver and a source impedance which is flat with frequency.
In the case of the anode driven tone stack, neither of these assumptions are true. Both are modified by the impedance to 0V vs frequency characteristic of the tone stack itself.

[catalin gramada]

Thanks. too many variables to take in consideration. For me cathode folower didn't works. As well triode fixed biased from B+ with large resistors dividers. Think fact is all depends how couplings conditioning to work the stages before and after and not only the tone unit for itself.

[catalin gramada]

Hallo. have little bit time to figure how my project looks like. Start with power supply sketch. Your comments are very welcome.please. Thanks. Catalin

For better understanding I figured 6 nodes decoupling stages I prefered to star carefully with no ground issues at all.

[KevinOConnor]

Hi Guys

catalin: You do not need the choke-filtered supply for the driver and preamp provided there is sufficient RC stages.This is more typically fed from the screen node, which is often choke filtered, so there is already clean DC being fed to the driver node RC. In your amp, with a UL output stage, the choke-filtered Va node is a good choice, but not entirely necessary if there are individual bias pots for the output tubes to allow hum balancing.

If you are using a concertina splitter, this needs as high a supply voltage as possible to have enough swing for the output stage. It should have a well-filtered supply node, so the second choke could be justified. In any case, the feed to the driver should be tapped from the choke-filtered Va node so all the nodes down the line can benefit from this expensive filter device. If you cascade the chokes, then this node should have series-connected caps to withstand the 570Vdc.

You are obviously planning on using 500V can-style caps everywhere. The dual-cap for the driver and second preamp stage introduces a compromise in the grounding. Personally, I would use snap-mount caps on a card, 300V or more each for the series ones and 500V units down the line. A bleeder down the line will help control how high the voltage gets without tubes in the amp - a zener string would assure against over voltage.

Your schematic shows an electrostatic screen between the primary and secondaries. Is this to be a custom PT? If it is custom, then you should save some weight and expense and have a cooler PT by abandoing the CTed winding and half-bridge, using instead a single winding and full-bridge rectifier.

Have fun

[catalin gramada]

Your schematic shows an electrostatic screen between the primary and secondaries. Is this to be a custom PT? If it is custom, then you should save some weight and expense and have a cooler PT by abandoing the CTed winding and half-bridge, using instead a single winding and full-bridge rectifier.

Have fun

Thanks for you reply. The supply was 'designed' for best linearity I can achieve without a regulator. PT is a regular Hammond 378cx (version of 278cx for Europe mains 230v) it have a electrostatic shield by default.

[nickb]

I can see you really didn't want to compromise here I don't see anything horribly wrong so what follows is just my take. We all have our own way of doing things.


The 100 ohm on the primary (low power /squishy mode ?) has the problem that it will lower the heater voltage which is a bad thing to do. It would be better to relocate it on secondary CT return.

The two fuses in the HT is overkill. A single one in the CT return will cover all likely cases other than a simultaneous failure of all four rectifiers.

I don't think you will have problem with ripple On the other hand all those chokes and big caps might have a problem with your budget (and your back). If you really want negligible ripple at a fraction of the cost you might consider a MOSFET type capacitor multiplier or regulator.

There is an problem in that when unloaded all the rails will get close to 600V allowing 5% transformer regulation. This will be an issue for the 500V caps. In normal operation there is no problem but suppose you lose the heater supply then you'll have the full 600V across them. The 18K resistor will limit the current so it's probably not a big deal but it's something you might want to think about.

Given you have a 50V tap on the 378CX for the bias I don't see why you need a separate transformer and full wave bridge. You didn't help yourself by having a 10K load on the supply -why so low? Also a 40V adjustment range seems a bit excessive and might be hard to use in practice The benefit is miniscule. Echos of busted budget again. Also, the bias supply is on the unswitched unfused side of the power.

I can never decide between fuse then switch or vice versa. It seems safer to have the fuse after the switch just in case some numbnut uses a screwdriver to scrape out bits of broken fuse. You know it can happen.

Good luck with the project I can see you are putting quite a bit of upfront effort into it and that is always a good thing.

Oh... are you having fun yet?

PS: Shame on me. I forgot the single most important thing of all - where is the safety earth?

[catalin gramada]

Hello. Thanks. Those 100 ohm/14 w is as important to keep inrush current under control at startup as well to drop some voltage into primary side for reasons you are talked about.It is a 'soft start switch'. Into a standby mode there are 35v dropped under it so a big power resistor is compulsory for 14w dissipation. I used a 50w aluminium housed (14w without radiator) so it is right on spot for few seconds I waiting before put the switch on.

...well I thought someone will ask me in the end. It is but is useless and for this reason I didn't figure. To explain myself in my country the mains for domestic use... simple haven't. The earth protection are reserved for mains plugs just for kitchen and bathroom. ...idiots.
And Yes fun all the way, I live for
Cheers. Catalin
P'S yes the 50v tap is in top limit but still enough to assure bias voltage. The main reason to not use it was to not share those common return. It cost me almost nothing and have plenty of room for a small toroidal. Also it will a small issue at start.up as I.m thinking. The bias supply is independent and will start at the first moment you plug the amp. The rest of supply is conditioned by the limiter resistor at start.

Man don.t expect to think as an engineer. I.m a truck driver and I write from a diesel station in my week.end break...