|
Post by thetragichero on Sept 2, 2018 12:24:29 GMT -5
edit 10/11/18: here is a link to a schematic I edited in red with notes and the values I changed to 1MB PDF Original Post: so i've been doing some minor tweaks here and there on my fender pawnshop excelsior because i dig it and it's fun i have a 5e3 kit build from years ago that's not working for some reason and i was wondering: could i pop the 5e3 output transformer into the excelsior? it's way beefier and would allow me a 4ohm tap for my killer 2x12 cabinet would it be louder since it's rated 20w instead of the stock 13w? this transformer is said to be a good replacement for the excelsior, and i've noticed the primary impedance is 6.6k instead of 8k for the other. not too high a price to pay if that would be better, but i'd like to get input on what i would look at to determine what would work and what wouldn't/might not here's the excelsior service diagramthanks in advance! edit: further reading tells me that with a cathode bias amp i wouldn't get 20w out of two 6v6
|
|
|
Post by reTrEaD on Sept 2, 2018 17:57:01 GMT -5
would it be louder since it's rated 20w instead of the stock 13w? Not necessarily. In fact you might get far less power out than the circuit was originally designed for if the primary impedance is different than the original transformer, even if the power rating is higher.
|
|
|
Post by sumgai on Sept 3, 2018 9:35:19 GMT -5
Yes, go for it.
The rating of a component is a measure of how much (current, voltage, power, etc.) can be safely passed through said component - it has nothing to do with what actually passes through it, that's up to the circuit designer, and how the component is used/misused/abused. 6.6KΩ is the recommended load for a pair of 6V6s in a Class AB push-pull configuration. 8KΩ is the nominal load value for a pair of 6L6s, but those should be fed with a significantly higher B+ voltage. (400VDC is the maximum rated voltage for the 6V6.) The Excelsior is rated at a lower output wattage due to a variety of factors, not the least of which is that the Harmonic Distortion rating is much lower, which translates to a clearer sound from the get-go.
The method of biasing has nothing to do with how much power can be derived from a given tube (or pair of tubes, etc.). Please note that the Deluxe amp, in its early incarnations dating back to the late 50s, had cathode biasing, and it was usually rated at 20 watts... though more realistically, it often came out at something a bit lower, like 16 to 18 watts. (The more widely known Deluxe Reverb, however, did have fixed bias, and it was rated at 22 watts for essentially the otherwise-same circuit. This is likely the source of the rumor about self-bias (cathode bias) having less power-delivering capability.)
The replacement transformer with multiple output taps is a nice touch.
HTH
sumgai
|
|
|
Post by thetragichero on Sept 3, 2018 10:10:33 GMT -5
i have changed the title as i might as well document what i'm doing with the amp for posterity's sake one of the first things i did was remove the pcb-mounted jacks and installed switchcraft jacks i added 100k resistors in parallel to R21 and R37 to slightly increase the upper bounds of the tremolo speed swapped the stock speaker with a sunn pa speaker i bought off ebay - to be honest i can't tell any difference in the sound
to do: change C1 and C3 coupling caps from .01uF to .1uF as this seems to be more fender-like (brownface princeton 6g2 is where i'm getting some ideas due to similarities)
add DPST switch to bypass sag resistors R22 and R30 so that it has a mushy/tight bass switch
add 47uF cap in parallel to C28 to increase bass tightness/decrease power supply 'sag'
on V1a keep plate resistor R5 as 220k but change cathode resistor R44 from 3.3k to 1.5k (or 2.2k) and cathode cap C18 from 10uF to 25uF. the 220k 1.5k 25uF combo is what i've found for V1a for a vox top boost channel... will be increasing the gain here but then to compensate for that... on V1b change plate resistor R7 from 220k to 100k, cathode resistor R6 stays 1.5k, cathode cap C7 from 10uF to 25uF (100k 1.5k 25uF looks like pretty standard fender values...)
i've read about changing power tub cathode bypass cap C32 from 100uF to 22uF to allow more "clarity, responsiveness, etc mojo terms." looks like that would cause a bit of bass cut with may be needed since the preamp is allowing more bass freq through due to larger cathod bypass caps. this is a maybe
remove the mic input and add a switched jack for a trem footswitch - this may take some more figurin` so that i'm not grounding either of the contacts of the switch to the chassis.
unlike how i usually operate, i will try to do these incrementally so i can get an idea of what each did/does and not just all at once. makes it easier to troubleshoot as well!
nonworking 5e3 build will eventually be cannibalized along with a partially-completed turret board 2204 preamp to make a 2204 preamp/5e3ish power amp head that i will make the cabinet for.... sometime
|
|
|
Post by thetragichero on Sept 3, 2018 17:02:06 GMT -5
pretty sure 5e3 wasn't working because there was a shorting jack used for one of the speaker outs. I'm betting that i was plugged into the other jack and all output was going to ground.....
regardless, i have started dissecting it and made some changes to the excelsior preamp. changed the plate resistors and cathode resistors and caps and the coupling caps. sounds great but the bass isn't where I'd like it. not exactly flabby, but could be more... defined. maybe the additional power supply caps will help when they come in. OT swap possibly tonight. this is fun!
|
|
|
Post by thetragichero on Sept 3, 2018 22:12:11 GMT -5
put in the 5e3 output transformer and it is LOUD (especially through my 4ohm sealed cabinet) didn't realize the size difference between the two until i put them side-by-side: imageuploadbass is thick and chunky but not flabby. waiting on a switch to try the sag resistor bypass, but for a day off i am satisfied
|
|
|
Post by sumgai on Sept 4, 2018 18:49:26 GMT -5
to do: change C1 and C3 coupling caps from .01uF to .1uF Good call, should improve the overall tone, but moreso in the bass region.
Not so fast there, Bobalooie. Those resistors were put there because the designer chose to dis-include a Standby switch. In response to that highly thoughtless ommission, R22 and R30 were put there to act as in-rush current limiters. They don't act on the circuit at most times outside of the initial in-rush during the first few moments of powering up. For you to bypass these two parts, thus allowing a full in-rush of current to the tubes during power-up, well, that's your call, of course. However, any guarantees on your tubes will read something like "for 30 seconds or 30 feet out the door, whichever comes first".
This is OK, it should work like it says on the tin. It's all cool here, it's nothing more than a matter of personal preference. Like I've said before, experimentation is the name of the game. You can't hurt anything (unless you create a solder bridge across a B+ line...), and you might find a new way of doing something that should be noted on these here innerwebs.
A bit more industrious than most mods. Try it, but I'm not holding my breath that it will have the desired effect. My gut feeling is that the bass region frequencies will sound dull, but that's just a thought, and I might not be correct.
Good plan! HTH sumgai
|
|
|
Post by thetragichero on Sept 4, 2018 19:49:54 GMT -5
Not so fast there, Bobalooie. Those resistors were put there because the designer chose to dis-include a Standby switch. In response to that highly thoughtless ommission, R22 and R30 were put there to act as in-rush current limiters. They don't act on the circuit at most times outside of the initial in-rush during the first few moments of powering up. For you to bypass these two parts, thus allowing a full in-rush of current to the tubes during power-up, well, that's your call, of course. However, any guarantees on your tubes will read something like "for 30 seconds or 30 feet out the door, whichever comes first". sumgai so long as i don't flip the sag resistor bypass switch until after its heated up i should be fine? power tubes are older than me made in USA rca so pretty stout anyway
|
|
pj
Apprentice Shielder
Posts: 48
Likes: 5
|
Post by pj on Sept 6, 2018 2:39:46 GMT -5
Cool project. These kind of amps are ripe for tuning to suit your own needs.
I would be careful about lowering the v1b plate to 100k. The concertina phase inverter behaves somewhat better when driven from a higher impedance. Also that kind of pi has less than unity gain so I would want to keep some of that extra amplification given the limited number of gain stages. I would perhaps insert some more series resistance before v1b limit blocking distortion and smooth the amount of high end clipping there.
If you wanted to maximise sensitivity and headroom you could experiment with the screens by switching from the 470ohms to 56/100 and a zener. This mod has been used to great effect on blues juniors and others and maximises output gain by regulating the screens far better.
Excited to see where you go with this.
|
|
pj
Apprentice Shielder
Posts: 48
Likes: 5
|
Post by pj on Sept 6, 2018 4:50:44 GMT -5
Also as regards sumgais spot on comment regards the inrush resistors, if you really wanted to play with sag you could switch in other resistors downstream of the inrush. With a two pole centre off switch you could set up two levels.... but... this amp has a full wave bridge rectifier and in my experience the benefits of the sag game are somewhat lost on this more efficient lower impedance design.
|
|
|
Post by thetragichero on Sept 6, 2018 8:15:08 GMT -5
thank you everyone for your input so far i think the amp sounds great besides the trem being so suble. next time i feel like fiddlin` i think will replace R15 with a 1M pot to control the depth of the tremolo. doing some reading on the power tube screen resistors. only mention of a zener diode is not in a good idea type light. also need to get a bit more rigorous into how the pentode/tetrode differences weigh into this idea (bjr is el84, excelsior is 6v6). as little more clean headroom is not a must but definitely something to ponder edit: re phase inverter, i have a 12DW7 (first triode 100mu/12ax7 second triode 17-20mu/12au7) i picked up for some reason or another that i may try in v2 since the first triode is tremolo circuit, second is phase inverter. according to this data sheet that is kinda what this tube was designed for okay enough delaying - time to leave for work!
|
|
|
Post by sumgai on Sept 6, 2018 11:23:12 GMT -5
so long as i don't flip the sag resistor bypass switch until after its heated up i should be fine? Potentially so. But I have a better idea... Since you're modding this puppy anyways, why not just do the right thing and install a standby switch? Put it series between the combined output of R22/R30 and R58, and you'll be golden. At that point, you can experiment with sag/anti-sag resistors all day, and be secure in the knowledge that you aren't harming/shorting the lifespan of your tubes.
Age has nothing to do with ability to withstand harsh conditions. Manufacturing quality has been all over the map ever since the early days of the Industrial Revolution. But I'll bet they sound mighty nice!
HTH
sumgai
EDIT:
I now realize that I've taken yet another shortcut (in my own mind), and not laid out the reasoning for a standby switch in the first place. Let me now correct that omission. (This is for those of you new to tubes in general, some readers will already be well-versed in the following discussion.)
In general, there are two parts to any tube (or valve, if you despise the American term). Regardless of the number of sections within a tube (like for instance a 12AX7 has two triodes, totally independent of each other), there will be an operating section, and a filament or heater section. Some of the early tubes combined them into one entity, but it was found early on that this design tended to fail much more quickly. They are still in use in some military gear, notably for their compact size, but in the general mainstream of things...
Now watch as we turn on the power to a tube. You'll recall that the B+ (the high voltage that can be dangerous) is going to the plate (or anode, as listed on some diagrams), but in point of fact, there's a circuit here, and if the plate were the only thing in the circuit, then there would not even be a circuit, would there. The other necessary circuit component, contained within the tube, is called the cathode. In short terms (very short!), the catchode emits electrons, and the plate receives them. To make that emission of electrons happen, the cathode must be hot, which gives rise to the Heater part of the tube. (Again, also called the filament.)
Now, if a cathode is "forced" to emit electrons whilst still cold, bad things happen. Essentially, without going into the chemical analysis of the whole chain of events, the cathode materials are shocked, and damaged a little bit. The thing will still work, but its lifespan has been foreshortened by some small fraction.
So how do we prevent this, and try to make our tubes live as long as possible? Simple - merely separate the two parts' "power-up" conditions. By that I mean, put them on separate switches. This lets us heat up the tubes first, then apply the high voltage after the tube reaches operating temperature.
The only othe way to derive some tube longevity is to use in-rush current limiting resistors. Talking about them is a topic best left for another day/post. <End of discussion>
Always remember - a happy tube is a good-sounding tube!
HTAH!
|
|
pj
Apprentice Shielder
Posts: 48
Likes: 5
|
Post by pj on Sept 6, 2018 13:21:01 GMT -5
only mention of a zener diode is not in a good idea type light. also need to get a bit more rigorous into how the pentode/tetrode differences weigh into this idea (bjr is el84, excelsior is 6v6). as little more clean headroom is not a must but definitely something to ponder I'll try and find some links for you to read up to give some more accurate background but the concept works irrelevant of tube type and is only really limited by how much voltage you need to drop and the b+ level which isn't an issue here. The screen is far more significant than is generally given credit in amplifier mod circles. The problem is that as the screen draws more current the voltage drop across the screen grid resistor gets larger, pushing the screen further from the plate, which has a compressive limiting effect. Fixing the screen with respect to the b+ (and having a smaller resistor for current limiting and stability purposes) ensures more linear gain throughout the swing at the grid. Hopefully that makes some sense, I may not have explained it well in between the kids fighting over tea. I will find you some links. Of course a good example of a half way house to the wonder of screen regulation is the way hiwatt particularly, and traynor (amongst others) would do the screens with a single dropping resistor from the b+ node and then smaller resistor to each screen. This ensures a lower overall voltage drop across the single larger resistor.
|
|
|
Post by thetragichero on Sept 8, 2018 21:54:55 GMT -5
of course the 12dw7 has the two triodes swapped from where I'd like it (12ax7 half would be in the phase inverter instead of tremolo). jj makes a 'reverse' 12dw7 so i ordered one. if i like what it does, i could always swap the wiring to v2 so i can plug in my nos mullard 12dw7 since the 12au7 triode can handle like10x the plate current of a 12ax7 (giving it the ability to drive the power section harder), I'm guessing that i would have to figure out what plate and cathode resistors would need to be changed to, eh? that would come from figuring out the load lines from the data sheet?
in other news, the first dpst switch i ordered from eBay (for sag resistor bypass) came... and was an spst switch. should work for standby switch. even looks like it matches mains switch
found four matching bf/sf fender-like knobs from the bin for when i add tone control and tremolo depth pots
played the amp for a little bit with me jazzmaster. sounds great with volume at 3 o'clock and into accordion input. almost but not quite chunky punk/metal tone. any further and it just compresses more. cleans up with volume knob. me likey
|
|
|
Post by thetragichero on Sept 9, 2018 9:59:11 GMT -5
online research for similar circuits suggests changing the cathode and plate resistors from 56k to22k (also suggested to make them2w instead of 1/2w, bias resistor somewhere in the range 1k-3.3k (currently it is 1k sho i shall leave it for now), possibly drop down the capacitor preceding it to 500-1000pf (it is currently 22nf) too prevent blocking distortion - going to leave that be as well unless i run into issues now to determine if i feel like heating up the iron after church....
|
|
|
Post by thetragichero on Sept 13, 2018 22:51:44 GMT -5
jj 'reverse 12dw7' came today and between that and the previous changes I've made (including swapping the 1M resistor from the tremolo circuit for a 1M depth pot) have got this baby howling! nowit starts to really get dirty around 6 or 7 on the knob. lower than that is loud and 'clean enough' i mainly used the accordion input and the makeshift tone control (just soldered the3 wires from bright/dark switch to the 3 pot luff) doesn't do all that much tried out my modded metal zone in front and i can even get a fairly usable metal tone out of this thing
still waiting on dpst switch to add a standby switch, sag resistor bypass switch, and maybe beef up the first power supply filter cap
this has been fun!
|
|
|
Post by thetragichero on Sept 15, 2018 15:07:51 GMT -5
fighting some blocking distortion. changed coupling cap to 4300pf (from 22nf). suggested was 500-1000pf but couldn't find one easily (at least a high voltage one) so i stuck in 4300pf and also added 1M grid stopper. still there. power tubes have 2k2 grid stoppers so i will look into increasing to 5k or 10k
since I'm using the 12au7 half of the preamp triode i decreased the cathode/plate resistors from 56k to 22k (didn't have any but i had some sweet 3w 33k resistors that i then added 56k in parallel). could this just be too much current I the phase inverter? info i read said 22k or47k, do maybe i go to just the 33k page/cathode?
besides that the only other annoyance is that i replaced the fake jewel lamp led with a standard fender-style lamp indicator and it doesn't seem to work. could the 150r current limiting resistor be the cause? it's powered from the heater tap on the pt and it's an incandescent so polarity shouldn't matter... only other thing i couldhonk of would be a bad bulb but i doubt that (suppose it's an easy thing to test next time i decide to mess with it
|
|
|
Post by thetragichero on Sept 15, 2018 16:56:17 GMT -5
another thing I'm reading is to change the 1M resistor to ground on the pi to ~68k to prevent this distortion
|
|
|
Post by thetragichero on Sept 15, 2018 21:02:51 GMT -5
just tested lamp with 9v battery and 250r resistor and I'm pretty sure that's why it wasn't working. blew a 7a fuse on heater tap trying to wire the lamp a different way first so now EVERYTHING is apart and i said the eff word more than i usually do these days. oh well
b+ is all the way up at 490v (schematic says 401v from first filter cap)I'm guessing that's why i started experiencing that blocking distortion
|
|
|
Post by thetragichero on Sept 15, 2018 22:26:02 GMT -5
fixed pilot light. added 82k resistor in parallel to 1M r25. didn't like it at all. will increase grid stoppers on power tubes another time
|
|
|
Post by thetragichero on Sept 16, 2018 7:58:48 GMT -5
real quick meter reading while my better half was getting ready for church has voltage from pin 7 of v2 at 358v. found a spreadsheet on some forum that calculates load line values for gain stages, cathodyne pi stages. plugging in my component values put me just below the 8v line that I've elsewhere read is my goal for 12au7 stage (in terrible for documenting sources... bad rigor on my part). this leads me too believe that the pi is doing what it needs and it's the power tubes that are causing the problem so i should increase the 2k2 grid resistors
this thread is as much for me to reason through and remember stuff as it is for posterity thanks for putting up with my ramblings/any thoughts are appreciated
|
|
|
Post by reTrEaD on Sept 16, 2018 11:07:43 GMT -5
real quick meter reading while my better half was getting ready for church has voltage from pin 7 of v2 at 358v. found a spreadsheet on some forum that calculates load line values for gain stages, cathodyne pi stages. plugging in my component values put me just below the 8v line that I've elsewhere read is my goal for 12au7 stage (in terrible for documenting sources... bad rigor on my part). this leads me too believe that the pi is doing what it needs and it's the power tubes that are causing the problem so i should increase the 2k2 grid resistors this thread is as much for me to reason through and remember stuff as it is for posterity thanks for putting up with my ramblings/ any thoughts are appreciatedUsing a thread as a public 'blog' on a project is totally legit and more than welcome. Thought: If you haven't done so already, I suggest bookmarking a link to Randall Aiken's website aikenamps.com/index.php/white-papers. He runs a boutique tube amp shop and imho hits the optimum balance of common sense with enough science in the explanations.
|
|
|
Post by thetragichero on Sept 16, 2018 16:35:20 GMT -5
that is one of the pages (along with valvewizard.co.uk (plus some pdfs off his book) and one particular tdpri forum commenter that i only recognize by his avatar that looks kinda like Lemmy) that i am getting the best info from
that particular page is leading me to change the grid resistors to some 1w 100k i found in the bin to see if it gets rid of the (unwanted) distortion. 100k is listed as having negligible effects on the (wanted) sound, and if i position them the right way i can test smaller values just by temporarily soldering parallel resistors. but this i shall test later
|
|
|
Post by thetragichero on Sept 16, 2018 20:44:56 GMT -5
man I am so glad everything has been done before. tremolo in the excelsior is the same basic tremolo circuit that's been used forever and apparently all i need to do to make it footswitchable is take a ts jack, wire r16 to tip and sleeve to ground. when the latching switch is closed the oscillator signal goes to ground, when opened it starts back up again! neat!
while i have the top part open (input->phase inverter, basically) i am trying to get the treble cut end of the tone pot to give just a *little* more cut. current values are 10nf and 22k which gives a corner frequency of something like 728hz (between f and f# above high e string). adding a 410 resistor should drop that down to 497 which is about b above the high E string. not a lot more cut but should be just enough to not have to turn the pot all the way when I'm pushing the front end with a boost pedal
|
|
|
Post by thetragichero on Sept 16, 2018 21:46:55 GMT -5
safety first, kids!
|
|
|
Post by thetragichero on Sept 16, 2018 23:16:08 GMT -5
changing grid resistors from 2k2 did not solve this blocking distortion it's most noticeable when the volume is lower and the guitar sounds clean will fiddle more tomorrow
|
|
|
Post by thetragichero on Sept 17, 2018 23:21:22 GMT -5
changed r17 and r26 to 33k. still that doubled, fuzzy distortion in the background while playing clean then socketed c23 after dropping to 500pf still didn't get rid of it. tried super small (68pf) to super large (big horkin .68uf) and none got rid of it. tried a resistor (i think220k) in series with c23. one way made the buzz worse, one killed volume (in parallel killed almost all output). tried bypassing pi grid stopper resistor along with placing 10m in parallel, tried10m and220k in parallel with r25. nothing so far has fixed the buzz, some things have created oscillations
also tried different preamp tubes to rule that out. actually got nice sound from v2 as a 12ax7 (felt like it gave more tele thump even though it was a bridge humbucker on the jazzmaster) but didn't leave it in too long since Rk Ra are 33k which i think is a little hot bias for 12ax7
blocking distortion/buzz while clean seems to have started when i added an extra 16uf electro in parallel with the first filter cap. will keep messing with it later. now is much overdue sleep
|
|
pj
Apprentice Shielder
Posts: 48
Likes: 5
|
Post by pj on Sept 18, 2018 16:52:37 GMT -5
Apologies for not following up on my earlier post my healths been a bit variable.
Sounds like you're having fun!
Have you got an edited schematic of where youre upto? I was a bit curious in a couple of places where you were referring to grid resistors and I wondered whether you meant grid stoppers or grid leak? Going from 68k to 1M for example at the pi, that resistor is not at ground, it is a grid leak but its setting the bias as well. So, I would look at that again. Not only would that place a load on the previous stage you'd have the bias of that stage very much different, and potentially deflecting much more than it should with grid variation.
Have you got a scope for tracing purposes? Do you trust the deluxe output tx after its previous problematic situation?
My first thought regards the increased B+ would be potentially that of a higher primary impedance in the upgraded tx. Where's your bias at now? If youre primary impedance is higher, and youre drawing less current, it could well be running too cold. Crossover distortion could well be the source of some of what youre hearing. Of course going the other way runs into trouble as sumgai pointed out re the capability of the supplies design limits.
If the blocking is there in the background when you're playing clean I would suspect that the problem was severe asymmetry caused by either the bias point of a stage being too far one way or the other, or interstage impedance being too low causing grid current to be drawn when the signal swings that way. I suppose that is essentially defined as blocking distortion but I think its useful to think of it the other way when fault finding because it has you looking for the source of the issue rather than trying to bandaid it. Of course it could be a nested combination of the above, hence the need for a step back and look at the overall picture.
Given the lack of gain from a cathodyne splitter I would argue that your chances of getting into blocking at the output are far less than other points mentioned above so I wouldnt focus too much on the stoppers there. You'd need to be hammering the pi before you got near any kind of grid conduction at which point the distortion would be more symmetrical. Having said that, if youve lowered the pi resistors considerably then there’s potential in that area.
I'll have another read back through and see if I can make more of a mental picture. Since my second cardiac arrest and associated oxygen starvation my ability to hold data short term has become cr@p so apologies again.
all the best, take care
|
|
|
Post by thetragichero on Sept 18, 2018 19:58:31 GMT -5
it would make a lot of sense for me to update the schematic with current values. make it easier than reading blocks of text. that may take me several days since it's in front of the computer work.
add far as i know, i didn't have problems with the deluxe ot besides the speaker being connected into the wrong jack (nc shorting Jack would'vebeen the right one) do no sound would come out the few times i would fire it up for <30 seconds over the past few years to see if it "still didn't work." stupid human tricks
|
|
|
Post by thetragichero on Sept 23, 2018 19:45:16 GMT -5
spent a couple days away from the amp/real life kept me busy. I've read a bunch of the 130-something page tdpri thread on the excelsior and many have had mechanical buzzes. so i plugged into extension cab and no buzz. plugged a head into the speaker in the combo and got buzz. so ours not electrical! going to take apart the cabinet and try just the bare speaker. still also take out every screw and wrap threads in teflon tape.
i am still editing the schematic to update with current values. if anything just for posterity
|
|