New Circuit, New Build, New Questions...

[Structo]

I've got to hand it to you flood.
I have no idea what is available in India but there probably is not the same availability of parts as there is elsewhere.

Are you a native born Indian citizen?
Just curious since your English is very good. I know that Britain had their fingers into a lot of countries back in the day so maybe that explains it.

Sounds like a cool design you are working on anyway.

[flood]

thanks for the compliments tom

india is certainly a curious case in terms of parts. with regards to availability, the entrepreneurial spirit rides high here, and you'll often find someone willing to take on an offbeat job. this, however, is not always a good thing, because the work ethic and attention to quality leaves a LOT to be desired in general.

i've been running around the electronics markets for years, so i know where to locate most of the stuff i want. sadly, you do need to import anything beyond a very basic quality. this includes film caps and pots. i bought my film caps in germany, so i have a lot of wima and ERO caps. pots.... well, i'm using the taiwanese alpha that you encounter so often in pedals. they've worked ok so far though.

so in general... it isn't really easy to find stuff - you do have to dig very hard. transformers, however, are not much of a problem - you have a lot of people here who can wind transformers at VERY reasonable rates with whatever data you give them. in fact, i believe that india does a lot of OEM transformer work - e.g., http://www.nuvotem.com/ are built by http://www.talemaindia.com/ (absolutely retarded that you can't buy them locally though!).

regarding my english - yep, india was a british colony between 166? and 1947. also, i grew up in south bombay, which bore most of england's legacy to bombay... the rest is mostly land reclaimed post-independence. my school was established by the english (see http://en.wikipedia.org/wiki/Cathedral_ ... non_School) and i even had a couple of british headmasters/mistresses as late as 1990!). so i'd go so far as to say that english really is my first language, with hindi coming second (although german would definitely equal english in terms of proficiency, since i lived and studied there for the longest time).

if there's one thing i'd REALLY like here - it would have to be that wire that you can bend to perfect 90° angles without it flopping over! the teflon wire i use for DC lines is very thick and fairly flexible but doesn't really do that 90° thing so well. i've seen some fantastic builds on these forums that i'd like to emulate. the other thing would have to be high-quality pots.

[flood]

IT'S ALIVE!!!

...but there remain a bunch of issues that need sorting out.

first off, it crackles like crazy when the guitar isn't plugged in. once that happens, it's fairly quiet.

the extra footswitchable boost does not work. not sure why.

i think both problems are related to a bad/badly grounded jack for the pentode channel input. also, since that was a very short length of wire, i skimped on the shielded. will change that tonight.

the most worrisome problem - the filaments show 7.3 VAC!!! the transformer was custom wound for 3.15-0-3.15 at 3A center tap referenced to output tube cathode. i can't figure this one out.

how does it sound? unlike any amp i've ever heard. the pentode channel sounded a bit gated at the beginning but after about an hour, it evened out a bit and wasn't nearly as harsh.

the parallel mode is definitely my favorite. the high-gain cascade mode is pretty unique, as far as i can tell.

other snags - the EQ doesn't seem to be doing, much, but i could be wrong about that.

working on the amp didn't zap me, but i did get a nice blast of what sounded like 1kHz feedback while testing for microphony.

voltages are nice - all under 300V. will be re-measuring them for documentation though.

tubes used: 2X ECC83, 1X EL84 (all BEL, india), 1x 6SJ7 (NOS sylvania)

i think this has been my most rewarding build so far. here's hoping i can iron out all the issues in the next couple of weeks.

[flood]

hey all,

here is the complete updated schematic with some major changes too.

now before i bombard you with questions yet again , here's an introduction/recap:

this amp has two input jacks - pentode and triode. it also has a 3 way switch (S3, preamp schematic) that goes between discrete, parallel and cascade modes (triode into pentode). i believe the tonal palette achieved by setting the pentode volume, triode volume and post EQ master volume is... versatile to say the least

now, the pentode jack is ONLY usable in discrete mode. for the other modes, the player must be connected to the triode input jack.

furthermore, there was an unused triode half that i decided to use for OTT gain (well, not really, since this amp is unusually voiced, it will never really do teh brootalz) that could be remotely switched. this led to the problem that signals would have to be mixed - so i tried to solve it the only way i knew, using large grid stopper and output resistors. (see preamp schem, V2B section).

i still have a couple of problems:

1. the footswitchable boost mode, centered around V2B in the preamp schematic, won't. all i get is noise. i devised it so that i could boost both inputs with minimal interaction - i'm beginning to see the shortcomings, though. if my theory is right, the 470k grid stopper is way too high and kills the incoming current. also, since shorting jacks are used and the unused input shorts to ground, there is an effective attenuation of 470k/235k or 2/3rd of the input signal. this could possibly be decreasing the signal to noise ratio, since the input swing is effectively too small. i think the output of the V2B stage is fine though.

when i disengage the relays, it goes back to bypass and works normally as desired. i checked to see if i'd grounded anything at both ends, don't see another culprit... is my idea moot? i considered decreasing the grid stoppers by a big value - say down to 68k - but that still gives us an attenuation factor of just above 50% and a sizeable current bleedoff to ground.

if it's any help: the 1M resistors are mounted directly onto the input jacks. the grid stoppers are (i know, not the best way to do things) a bit far apart from the grid. the grid leak is very close to the grid.

2. the EQ doesn't seem to be doing anything, at least as far as the mids and bass controls are concerned. the deep switch, however, is wicked, and the treble shift switch is being reworked. it seems i wired up the treble pot backward. also, there is a huge increase/cut in volume directly proportional to the treble cut/boost. i'm going to check it anyway, but what's your take: grave wiring error (a given), circuit design error, or the fact that i'm playing the high strings of a six string bass?

3. i'm still looking for an ultra-simple line out circuit (with a volume control) that doesn't fry the input of the following stage (audio interface, line level). i'm using a 10 ohm 50W resistor in an aluminum heat sink as a dummy load. what woould be the ideal voltage divider? i was thinking 100k//100nF + 10k with a 100k pot acting as a volume control.

on the whole, i'm relieved it's performing as intended and sounding jes' fine... need these kinks worked out. and i need my guitar back. never buying a floyd rose equipped guitar again; i figured i'm never going to be steve vai. also, i still have to try it with 6v6, 6L6, 6K6, 6F6, EL34/6CA7.

thanks to everyone who made this project possible so far! just a little bit more to go. please keep those great inputs coming! going to put it through the acid test and pop an EL34 in there now... i hope nothing bad happens.

[Structo]

Maybe you can borrow the boost circuit from the Dumble ODS?

It simply lifts the tone stack from the circuit.
Since it is a plate loaded tone stack it is fairly lossy, so when you lift it you get a good boost.

The ODS just breaks the connection between the bass and treble pot for the boost.

Another thing to try is:
Instead of lifting the entire tone stack, try switching in a 33k-47k resistor with a .001-.005 cap in parallel off the midrange tab to ground. It boosts gain and volume without deactivating the tone controls.

EDIT:
Just reviewed your schematic and yours is a cathode follower tone stack so I guess my idea won't work.

[flood]

thanks for chiming in nonetheless. i will keep that in mind if i design an amp with a plate-followed EQ (my 4W blues express could profit...)

i think i have found a solution, tube purists will be greatly miffed but i'm fine with the idea. i'm going to implement a signal splitter within the boost circuit.

it's simple and FET based: http://www.muzique.com/lab/splitter.htm

let's hope it works....

EDIT: i meant buffer! not splitter! one buffer per jack input.

[flood]

ok, update and quick question - i've decided to bite the bullet and uncomplicate things a bit. how much MORE noise am i risking if i do NOT tie the heater filament center tap to the cathode of the output tube, tying it directly to ground instead?

i really need to run a bunch of DC circuits off the heater AC; with the FET buffer coming in, it would NEED to be tied to the main ground... hence my question.

am planning to add an additional bridge recto and a 7806 - a bit low, i know, but should be fine for a small signal buffer.

[Structo]

I have a couple cathode biased amps that have the heater center tap to ground.
I don't notice much hum at all.

I meant to try that trick of grounding to the cathode but never got around to it.

That is a very complicated build so I guess anything is possible.

[flood]

ok, newest update: boost circuit disposed of entirely. it's overcomplicating things, causing too many problems and slowing down this build way too much. so bye bye boost. rewiring the amp now.


does anyone have any recommendations regarding the pentode? i am curretnly using a 4M7 resistor between B+ and grid with the cathode tied directly to ground, was considering dropping that to 1M-2.2M with 1k5/1u (up to 25u) on the cathode.

how will this change things? will there be an increase or decrease in gain? i think that it would lead to some tonal improvement. i still want to maintain the high gain i'm getting right now, and that fat pentode tone...

[flood]

so......

some more changes took place:

- boost and relays disposed of altogether! but we now have an unused triode half . maybe i'll parallel the halves.
- added the 1.5k//47µ on the cathode of the 6SJ7, dropped screen grid resistor to 2M 1W, and coupling cap from pentode plate reduced from 360nF to 10nF. it doesn't sound as fat anymore, but that weird blocking distortion is no more. it still sounds really good - fat, present and crunchy.

and it's nearing completion. here are a couple of pics for you lovely people:

[IMG:800:600]http://i38.tinypic.com/mbtulw.jpg[/img]

lousy guts

[img:800:600]http://i33.tinypic.com/2i0c07c.jpg[/img]

i do not like my wiring on this amp. granted, it's less of a mess than my 2061x, but it's still nasty. also, in retrospect, i think putting the PSU on a PCB was a bad idea, primarily for service reasons. it got extremely difficult to make turret board stay in there. maybe i'll change it nonetheless.

how does the amp sound? fantastic so far; extremely versatile, and tonally very complex. still need to install a couple of things, like the 6 way dark switch and the LEDs. like i said, the PCB might just be replaced with turret board or something else, if i have to remove the transformer. the wiring will be neatened up a bit too.

hopefully my last questions for this build:

1. however i do have a BIG problem - the PT i'm using specs the filament windings as 3.15-0-3.15. with all tubes installed like in the first picture, i get 7.3 VAC across pins 2 and 3 of the unused octal socket! measuring each pin to ground gave me about 3.4 VAC.

did the winder mess up the windings? is the impedance too low for the load? i am a bit worried about this because i think the extra volt will make the buyer go through tubes faster than popeye going through that can of greens.... i might have to send the tranny for rewinding

2. would there be a significant increase in fatness if i increase the 10n cap on the 6SJ7 plate to a multiple thereof up to 100n? from what i know plate impedance is very high in pentodes, so a small coupling cap can be used with little loss in low-end bandwidth.

please chime in, as always thanks!

[Rick]

When I run into this, (the 115V,117V primary PTs are prone to this) I usually try 2 1N4007s parallel forward and reversed (opposite directions) in one line of the filament wire off the supply. This should drop about .7V and get you back to a safe range, yours is unusually high, it may need more drop than that. 2 pairs would give about 1.4 V drop. Try it and see if you can get it around 6-6.3 VAC. Good luck!

[Structo]

You should have your tubes heaters wired in parallel and you measure the ac voltage across the two wires. That is what should be 6.3v.
If you are above 6.8v your tubes are not going to last very long.

How is the B+?

You have to have the tubes in when measuring heater voltage because depending on you PT, the voltage sometimes is pulled down if the design of the PT isn't very stiff.

You can use high power resistors to drop the voltage, use silicon diodes like Rick suggests or zener diodes but make sure they can handle the power or current of the heater winding.

[flood]

hey guys,

thanks for the tip. the idea of using diodes never occured to me, i must admit... i think it's worth looking into, particularly since rewiring the amp at this stage would be a nightmare of sorts, and the transformer winder ios in a different city altogether. the last time i checked the B+, it was around 300 on the plate of the power tube, but that was a while ago. i'll measure it again - i need to measure the node voltages anyway for documentation.

i think the PT could in fact be VERY stiff - it's rated 3.15-0-3.15 3A and all the tubes (2x 12ax7, 1x6sj7, 1x EL84) were in while measuring the heater voltages.

how about using a zener diode on the center tap to ground? i wonder if i can find one rated 1V 5W... that could give me 6.3 bang-on.

[Rick]

I should have read your post more carefully. You should put in all the tubes and measure across the two feed wires With them all lit up you may be much closer to the correct voltage than the way you described measuring them. Usually fils under too high voltage will visibly be too bright and older, weaker tubes will often go poof in a hurry! That is how I first discovered it, the single pair of 1N4007s will generally drop the voltage into a useable level. Running fils at 6VAC will not hurt anything that I know of and will probably improve longevity. If they read close to 6.3VAC with all tubes in you won't need to do anything.

[flood]

hey rick, thanks again. would UF4007 work as well? also, can i put the single pair on the centre tap instead of one of the sides? the reason i ask is because i would imagine that putting it onto one of the heater leads would make the voltage asymmetrical across the filaments.

....or am i simply nit-picking?