Marshall Class 5 cooks el84's
- Thread starter CJ
- Start date
Help Support GroupDIY Audio Forum:
7 mods, see white arrows
change cathode bias R to 220 ohms, was 150 ohms,
up screen drop from 1 K to 1.8 K
47 pf from 6BQ5 grid to gnd, kills parasitics,
treble cap was 470 pf, too mushy, change to 120 pf,
too much woof, change cathode bypass cap on sage 1 from 10 uf to .22 uf
change driver stage plate R from 220K to 100K,
and rewind transformer ,
change cathode bias R to 220 ohms, was 150 ohms,
up screen drop from 1 K to 1.8 K
47 pf from 6BQ5 grid to gnd, kills parasitics,
treble cap was 470 pf, too mushy, change to 120 pf,
too much woof, change cathode bypass cap on sage 1 from 10 uf to .22 uf
change driver stage plate R from 220K to 100K,
and rewind transformer ,
Attachments
mod info
http://www.mylespaul.com/threads/some-mods-for-marshall-class-5.123816/
http://www.tonegeek.com/musicgear/amps/marshall-class5mods.php
stock OPT was only 4 Henries, offshore job, 68 EI (actually metric but close to 68 EI) wound pri-sec-pri but had very different turns on he two primary pies,)
we upped pri to 3200 turns instead of 2000, used bi-fi #26 to lay flat for 157 turns instead of 97, and reduced gap thickness,
http://www.mylespaul.com/threads/some-mods-for-marshall-class-5.123816/
http://www.tonegeek.com/musicgear/amps/marshall-class5mods.php
stock OPT was only 4 Henries, offshore job, 68 EI (actually metric but close to 68 EI) wound pri-sec-pri but had very different turns on he two primary pies,)
we upped pri to 3200 turns instead of 2000, used bi-fi #26 to lay flat for 157 turns instead of 97, and reduced gap thickness,
Attachments
Jaysus ,
a total redesign then ,Ive worked on the same amp twice in a few years ,I didnt get so invloved trying to fix the designers funk ups though ,anything valve with silicon rects without a standby Im suspicious of at this stage ,could it be some kind of marketing decision to try and sell more spare tubes ? Ive seen alot of EL84 amps that turn even bad ass russian 7189's into a bag of rattley bones in no time .I hear the designers cry,put in a standby ,it costs 2 dollars ,the money mean men say nay nay nay ,
Marketing is the devil ,and we are the headlights on the black diamond train hell,all aboard.......
a total redesign then ,Ive worked on the same amp twice in a few years ,I didnt get so invloved trying to fix the designers funk ups though ,anything valve with silicon rects without a standby Im suspicious of at this stage ,could it be some kind of marketing decision to try and sell more spare tubes ? Ive seen alot of EL84 amps that turn even bad ass russian 7189's into a bag of rattley bones in no time .I hear the designers cry,put in a standby ,it costs 2 dollars ,the money mean men say nay nay nay ,
Marketing is the devil ,and we are the headlights on the black diamond train hell,all aboard.......
exactly. check his out:
we spotted a 38 K plate resistance value in the tube book for EL84 and wondered, "what is this?"
so just for kicks, (or by ignorance) , figuring it best to match that load with the OPT for best fidelity, we wound a 4000T to 92 turn OPT, 43:1 (16 ohm speak) ratio, stuck it in there, wow! fuzz box city, nice rasp like we have never heard but low watts, so we plugged it into the software to see what gives and his is what we got, might make a permanent amp like this for an effects box, sounded like Billy G on crack,
had to rewind so customer would not flip out and spread rumors around town,
notice the flat section of he transfer plot:
we spotted a 38 K plate resistance value in the tube book for EL84 and wondered, "what is this?"
so just for kicks, (or by ignorance) , figuring it best to match that load with the OPT for best fidelity, we wound a 4000T to 92 turn OPT, 43:1 (16 ohm speak) ratio, stuck it in there, wow! fuzz box city, nice rasp like we have never heard but low watts, so we plugged it into the software to see what gives and his is what we got, might make a permanent amp like this for an effects box, sounded like Billy G on crack,
had to rewind so customer would not flip out and spread rumors around town,
notice the flat section of he transfer plot:
Attachments
I probably should be posting this in brewery for all the unenlightened souls ,but anyway have a listen ,makes me feel good about the life and road Ive chosen ,
https://www.youtube.com/watch?v=xQbQeGKscOE
Well above and beyond the call of duty as usual CJ .
https://www.youtube.com/watch?v=xQbQeGKscOE
Well above and beyond the call of duty as usual CJ .
> figuring it best to match that load with the OPT for best fidelity,
No.
That gives good (not best) Power Gain. But we can get gain with little tubes.
It gives terrible Power Output. And you will struggle to get enough Inductance to reach 20Hz. (You can probably make 82Hz OK.)
Why? A pentode has *TWO* plate resistances. Above the knee it tends to 38K. Below the knee it tends to 400 Ohms. Pull a graph and draw the slant-lines.
As a general thing, the Maximum Power will be near the Geometric Mean of the 38K and 400r. Like 4K.
As a more work-bench thing: the ratio of V and I is very near a good load impedance. 260Vpk and 35mAp is like 7.4K. On a soft-knee EL84, go either way, the power falls slightly but the harmonic profile against level ("loud tone") changes significantly. Low-loading gives "better" power number and many classic amps go this way. 6.4K is totally ballpark. Considering we don't know what "16 Ohm speaker" really is (100r at 90Hz and 5KHz), it is exactly right.
No.
That gives good (not best) Power Gain. But we can get gain with little tubes.
It gives terrible Power Output. And you will struggle to get enough Inductance to reach 20Hz. (You can probably make 82Hz OK.)
Why? A pentode has *TWO* plate resistances. Above the knee it tends to 38K. Below the knee it tends to 400 Ohms. Pull a graph and draw the slant-lines.
As a general thing, the Maximum Power will be near the Geometric Mean of the 38K and 400r. Like 4K.
As a more work-bench thing: the ratio of V and I is very near a good load impedance. 260Vpk and 35mAp is like 7.4K. On a soft-knee EL84, go either way, the power falls slightly but the harmonic profile against level ("loud tone") changes significantly. Low-loading gives "better" power number and many classic amps go this way. 6.4K is totally ballpark. Considering we don't know what "16 Ohm speaker" really is (100r at 90Hz and 5KHz), it is exactly right.
thanks PRR! makes sense now. my electronics class was before pentode tubes came out,
Rp is just that delta V vs delta I along the flat part of the tube graph, kind of like a Collector graph for a transistor,
for max pwr, it looks like we should pick a voltage in the middle of what the tube can handle to get max swing, find the bias voltage along the max pwr line, strike a tangent which will be our loadline, the slope representing out transformer Z, this tangent slope will go up as we use lower B+ and go down for higher B+, xfmr Z will go down for low volts, and up for hi volts, ok then,
we want to go vertical and change dynamic plate resistance by changing bias, 38K Rp is just what happens to current when we change supply voltage, don't really care about that too much,
have to use a tangent so we do not cross over max pwr line,
oh wow, Inside the Vacuum Tube, 1945, should have read this first>
https://ia600702.us.archive.org/30/items/InsideTheVacuumTube/InsideTheVacuumTube.pdf
Rp is just that delta V vs delta I along the flat part of the tube graph, kind of like a Collector graph for a transistor,
for max pwr, it looks like we should pick a voltage in the middle of what the tube can handle to get max swing, find the bias voltage along the max pwr line, strike a tangent which will be our loadline, the slope representing out transformer Z, this tangent slope will go up as we use lower B+ and go down for higher B+, xfmr Z will go down for low volts, and up for hi volts, ok then,
we want to go vertical and change dynamic plate resistance by changing bias, 38K Rp is just what happens to current when we change supply voltage, don't really care about that too much,
have to use a tangent so we do not cross over max pwr line,
oh wow, Inside the Vacuum Tube, 1945, should have read this first>
https://ia600702.us.archive.org/30/items/InsideTheVacuumTube/InsideTheVacuumTube.pdf
Attachments
> Rp is just that delta V vs delta I along the flat part
Right. But the curve has two flattish parts, with a knee in the middle.
Plotting below the knee, you find a couple hundred ohms. See attached, purple doty lines.
This kink is why a pentode can swing wide with low plate loss.
The "best power load" is much more than the 200 Ohms, but much less than than the 39K.
Because these are so far apart (200:1), the optimum is very broad. This tube can probably be worked into 1K or 10K.
The next optimization is the plate ratings for power, current, power. If it is sold as an "audio" tube, there will be some good balance available. (Always read the "sales brochure", the tube datasheet.)
Or if existing design gives you a certain supply voltage, work back from there.
The "Z=V/I" math assumes the pentode has low losses, and about equal at both ends; also that THD may be guitar-amp level (5% or more). Most spec-sheets hit this +/-20%, showing it is not-wrong. Triodes are a little more complicated. "Low THD" requires running richer and avoiding low current.
Right. But the curve has two flattish parts, with a knee in the middle.
Plotting below the knee, you find a couple hundred ohms. See attached, purple doty lines.
This kink is why a pentode can swing wide with low plate loss.
The "best power load" is much more than the 200 Ohms, but much less than than the 39K.
Because these are so far apart (200:1), the optimum is very broad. This tube can probably be worked into 1K or 10K.
The next optimization is the plate ratings for power, current, power. If it is sold as an "audio" tube, there will be some good balance available. (Always read the "sales brochure", the tube datasheet.)
Or if existing design gives you a certain supply voltage, work back from there.
The "Z=V/I" math assumes the pentode has low losses, and about equal at both ends; also that THD may be guitar-amp level (5% or more). Most spec-sheets hit this +/-20%, showing it is not-wrong. Triodes are a little more complicated. "Low THD" requires running richer and avoiding low current.
