PDA

View Full Version : Amp project



Rockfactor
15-08-2017, 07:21 PM
In the late 90's I had to have a Marshall amplifier so grabbed a JCM900, 50W dual reverb with 2 x 12" speakers. Never really liked it too much and couldn't get the sounds I need.

I now use a modeller straight into the FOH desk 99.5% of the time. At two venues I plug the modeller into the fx return of the amp bypassing the preamp straight into the power amp section with the power switch set to 25W and the volume set to 2 1/2, so barely ticking over.

This now has me thinking, I'm not attached to the amp so maybe I could rebuild it into something more suitable. As an electrician come electrical engineer I know my way around electricity safely and have a fairly rudimentary understanding of tube amps so am contemplating rebuilding it as a 15W - 18W?

This would allow me to build the power amp section so I could use it with my modeller and then build some preamp modules to suit the sounds I'm looking for.

I replaced the 5881 valves with EL34s previously so I have a lead in question. As stated my understanding of valve amps is rudimentary at best but if all else stayed the same and I halved the amplifier voltage, this in turn would have the current so it would seem this would result in 1/4 the power output. Is this correct? I'm just contemplating if I can use the EL34 in a lower wattage amp or do I investigate using EL84?

Anyway I'm only at the maybe/what if stage so any thoughts, suggestions or harebrained schemes welcome.

Robbo


Sent from my iPad using Tapatalk

Marcel
15-08-2017, 09:58 PM
If tubes worked in a perfectly linear way then yes your theory is correct, however tubes have a great reluctance to work that way so if your were to try it then you'd probably be hugely disappointed with the resulting sounds out of your amp.

There are ways of reducing output power and/or volume that limited effect on the tone/character of the amp, and there are many more different modifications that can set the tone/character of an amp.

https://robrobinette.com/How_the_Marshall_JCM800_Works.htm is a good start on getting an understanding of the internal workings of Marshall amps, and it has links to popular modifications. At a minimum it offers plenty of useful food for thought...

DrNomis_44
16-08-2017, 12:10 AM
In the late 90's I had to have a Marshall amplifier so grabbed a JCM900, 50W dual reverb with 2 x 12" speakers. Never really liked it too much and couldn't get the sounds I need.

I now use a modeller straight into the FOH desk 99.5% of the time. At two venues I plug the modeller into the fx return of the amp bypassing the preamp straight into the power amp section with the power switch set to 25W and the volume set to 2 1/2, so barely ticking over.

This now has me thinking, I'm not attached to the amp so maybe I could rebuild it into something more suitable. As an electrician come electrical engineer I know my way around electricity safely and have a fairly rudimentary understanding of tube amps so am contemplating rebuilding it as a 15W - 18W?

This would allow me to build the power amp section so I could use it with my modeller and then build some preamp modules to suit the sounds I'm looking for.

I replaced the 5881 valves with EL34s previously so I have a lead in question. As stated my understanding of valve amps is rudimentary at best but if all else stayed the same and I halved the amplifier voltage, this in turn would have the current so it would seem this would result in 1/4 the power output. Is this correct? I'm just contemplating if I can use the EL34 in a lower wattage amp or do I investigate using EL84?

Anyway I'm only at the maybe/what if stage so any thoughts, suggestions or harebrained schemes welcome.

Robbo




Sent from my iPad using Tapatalk


If you're aiming to reduce the output power of the amp, you'd probably get better results if you reduced the number of output tubes rather than reduce the voltage, as Marcel said, tubes don't behave in a linear fashion like transistors do, you need to have the anodes of the tubes at a high positive potential with respect to the cathodes in order for the electrons boiling-off the cathodes to overcome the space-charge between the control-grid and cathode, an electron carries a negative-charge, and while the space-charge exists between the cathode and control-grid, there's a negative-charge on the control-grid which repels the electrons back to the cathode, it's this space-charge between the control-grid and the cathode which is responsible for the non-linearity inherent in all tubes that are used to amplify a signal, the positive potential on the anode attracts the negatively-charged electrons and a current flows from the cathode to the anode as the electrons overcome the space-charge (depending on which current-flow convention you're using), but, the current flow is not linearly proportional to the potential on the anode because of the space-charge limiting the current flow at low anode potentials.

@ Marcel, Feel free to check and see if my tube theory is correct.

Marcel
16-08-2017, 10:59 AM
Your pretty right there Doc....

Each tube has design specifications... They specify the ideal voltages that should be on the anode, cathode and grids for it to operate in a linear or other mode for which it was designed to work. For any given tube there is quite a bit of lee-way in theses voltages, but the variance is typically limited to plus or minus 30% from the optimum values whatever thy may be. That's not to say a tube won't work at greater variances, however when at those greater variances we are less likely to have linear or other characteristics that are useful and more likely to have undesirable side effects forced upon us.

In my AC15 build my EL84 anode voltages are a good 25% above the maximum rated voltage. The tubes I have are accepting of that higher voltage but for the tube to remain within its safe operating curve changes in the circuit needed to be made. Also due to the higher voltages everywhere because I used diode rectifiers instead of a tube rectifier I have V2b at the maximum of its heater-cathode voltage limit so changes will be needed there too.

Presuming our heater voltages remain at their ideal setting when reducing all the other voltages we move the operation of the circuit to the lower limits of the linear operating curve, the part of the curve where it really starts getting curvy. Yes it is operating in a semi-linear fashion however undesirable distortion elements caused by the curvy linearity start becoming significant and will "colour" the amp in very odd and possibly disturbing ways. No doubt many have experienced an amp with a failing rectifier tube which gives a perfect example of this scenario.

Back to Robbo's amp, maybe a few simple mods might be better to start with....
(Obligatory WARNING Note!!! In tube amps the HIGH VOLTAGES CAN KILL.!!! And the mods I'm about to suggest are right up in there with the deadliest for anyone considering them!!)
- A 'Post Phase Inverter Volume control' or PPI volume is probably one of the simplest as it allows the volume to be pulled back substantially without colouring the amp thus giving the opportunity to drive the amps input harder than it otherwise would be. Essentially it is replacing the two 220K (typical value) grid bias resistors with a dual gang 250k volume control referenced to the bias voltage (Not earth !!!). A pair of 2M2 'safety' resistors are included to the grids to ensure bias voltage can always get applied to the output tubes grids should the new dual gang volume pot ever go faulty.
- A switch on the EL34 output tube screens. By placing an appropriately rated DPDT switch on the screens to be able to select between the full HT power supply voltage or that particular tubes anode voltage is quite an effective high/low power switch. A side effect is that in low power mode (-3dB or 50% less power) when the screens are connected via their resistors to the anode voltage the EL34 output tubes effectively becomes a triode and will sound quite unique and different to the full power mode.

Plenty more to think about...

Rockfactor
19-08-2017, 09:48 AM
Thanks, this all,the good discussion I was hoping to start. As I said I have a basic understanding, just trying to come up with a good, usable and flexible amp. The plan is to see if I can reuse most of the stuff from the JCM900 in some way to keep costs down. This may not be possible but that's the starting plan anyway.
The JCM900 has the 25/50W switch that changes from pentode to triode mode (I use the 25W mode) but can still only run the amp at 2.5 and it's still very loud. The JCM900 is a pcb based amp so I am contemplating rebuilding the power amp section using turrets etc so I can modify a bit more readily (the PPI volume sounds like a good idea) for use when I am using the modeller as the preamp.
Then after that I could mod / build preamps to try and get the sound I'm chasing when not using the modeller.
It's as much about learning / experimenting as it is about the final result.


Sent from my iPad using Tapatalk

Marcel
19-08-2017, 03:11 PM
I found on a Chinese E-Bay site some 250k dual gang pots for about $3 each so I got a half dozen. When they turned up I found thy were linear taper 'balance' pots with a centre dent to indicate mid travel. Turns out they are very handy for PPI volume control use as the dent is effectively the half volume mark.
I installed one as a PPI volume on my 2204 build on the back of the chassis. Very handy as I don't need to look at it, I can reach around and with the dent go just by feel, and know where I'm at with that control. For me it defeated any need for a 25/50W switch on that amp.

One sad thing about the PPI volume control is unless you are set to maximum you do loose any ability to drive the power amp stage (and hence your speakers) into those over driven tones popular in harder rock type songs. At full PPI volume there is no problem, but at lower settings you become far more reliant on pedals and the first two tubes to do all your tone shaping which some people find is okay and others find is untenable. If you never drive the output stage or your speakers hard then a PPI volume control is a great colourless way to go to get manageable lower volume levels from your rig...

DrNomis_44
26-07-2021, 12:07 PM
Thanks, this all,the good discussion I was hoping to start. As I said I have a basic understanding, just trying to come up with a good, usable and flexible amp. The plan is to see if I can reuse most of the stuff from the JCM900 in some way to keep costs down. This may not be possible but that's the starting plan anyway.
The JCM900 has the 25/50W switch that changes from pentode to triode mode (I use the 25W mode) but can still only run the amp at 2.5 and it's still very loud. The JCM900 is a pcb based amp so I am contemplating rebuilding the power amp section using turrets etc so I can modify a bit more readily (the PPI volume sounds like a good idea) for use when I am using the modeller as the preamp.
Then after that I could mod / build preamps to try and get the sound I'm chasing when not using the modeller.
It's as much about learning / experimenting as it is about the final result.


Sent from my iPad using Tapatalk


I seem to remember looking at the JCM900 schematics, and noting that the hi-gain channels used an unusual silicon-diode bridge-rectifier configuration to generate clipping-distortion in the hi-gain preamp, this can sound harsh compared to the natural-distortion you get by overdriving 12AX7 tubes, Mike Soldano took a different approach when designing his Soldano SLO 100 amp, using cascaded 12AX7 gain stages.

Simon Barden
26-07-2021, 03:57 PM
There's diode clipping in the JCM800 as well. Now an established part of the modern Marshall sound. Blackstar also use it.

DrNomis_44
26-07-2021, 06:03 PM
There's diode clipping in the JCM800 as well. Now an established part of the modern Marshall sound. Blackstar also use it.


You're absolutely right, I seem to remember a good friend of mine owning a dual-channel Marshall JCM800, and when I looked at the schematics it did indeed have diode-clipping in it, the Marshall Silver Jubilee (the one that Slash from Guns n Roses used) also uses diode clipping.

Simon Barden
26-07-2021, 06:15 PM
I had a Silver Jubilee 50/25W mini-head (about the same size as the Sound City head), though mine was in factory-fitted black as a custom order to match the 4x10 cabs I had. Part-exchanged it for peanuts as I wasn't using it; now wish I hadn't.