12AX7 as a reverb driver

[Patrik]

Hi guys,
I am building an amp that is currently designed to use three 12AX7 and one 12AT7 as a reverb driver (typical Blackface reverb). Due to design tho one half of 12AX7 is unused, therefore I was thinking if I could replace 12AT7 in the reverb with one half of the 12AX7. I think it should be possible, but I have never seen it done before, so any opinions are welcome.
Thanks in advance.

[sluckey]

It's done quite often. Don't expect strong reverb though. Read this for more info...

     https://el34world.com/Forum/index.php?topic=7957.0

PS... Nothing wrong with having an unused triode.

[Patrik]

I could left one triode unused but I would like to utilize all the tubes I have in my dissposition. And problem with 12AT7 I have is that it draws quite some current for the anode, some 10mA but 12AX7 only draws 1.2mA, and I don't want to use bigger transformer, becouse with that I would rise B+. I don't mind weaker reverb as I just use it to put some more colour in to my playing.

[sluckey]

...And problem with 12AT7 I have is that it draws quite some current for the anode, some 10mA...

All the blackface Fender 12AT7 reverb drivers use slightly less than 4mA.

I would suggest using a 12DW7 if you want to stay with the one triode reverb driver. Use the low gain triode (equiv. to 12AU7) for the driver and the high gain triode (equiv. to 12AX7) elsewhere.

Tell us more about your projec. Providing a schematic helps.

[Patrik]

The schematics I made originally for this amp used two channels, tweed style, but after deciding to only use one channel I was left with an unused tridode of 12AX7 so I was thinking that maybe I could replace 12AT7 with that triode due to matching gain figure, 100 for each triode of 12AX7 and 44 for 12AT7 and since 12AT7 is wired in parallel output should have a gain index of about 90.
You said that in blackface amps 12AT7 only draws 4mA so in that case I couls use it. And then put in the amp another preamp stage.
You can see the two channel schematics of the preamp bellow. Techically I could then put another gain stage between tone stack and gain controll.

[Latole]

Nice sketch Patrik .

[Patrik]

Thanks for compliment Latole. Bellow is a photo of revised amplifier, making it single channel and more delay-friendly.

[2deaf]

Thanks for compliment Latole. Bellow is a photo of revised amplifier, making it single channel and more delay-friendly.

Too much dry signal compared to wet signal.  Reverb control will kill all signal when turned clear off.  V4 doesn't have a grid reference when something is inserted into FX Send.  FX Send output impedance is too high when the reverb control is at higher settings.  Signal at FX Send is way too large at typical settings.



 

[Patrik]

Thank you, I am still quite new to the world of amplifier design so things like the ones you mentioned just aren't apparent at the first glance. Thank you a lot. Original BF amps have 3.3M reverb mixing resistor so putting that in would help to solve the issue of too much dry signal? Or another option would probably be, to put tone stack in the same place as is the gain control and move one gain stage post fx as a buffer of sorts.

[DummyLoad]

+1 more problem with bias tap shown as drawn, if it's the atypical 50V tap, then the 100K in front of the bias rectifier diode is moot. your B- will be so low, you'll cook the output tubes and possibly other things. 100R-1k is more suitable value for that part. 


--pete 

[2deaf]

Original BF amps have 3.3M reverb mixing resistor so putting that in would help to solve the issue of too much dry signal?

It will reduce the ratio of dry signal to wet signal, but then you would need a 250K potentiometer at the FX Send in order to get enough signal.  A 250K FX Send pot is unacceptable because an impedance this high will form a high-cut filter with your patch cord.

[2deaf]

+1 more problem with bias tap shown as drawn, if it's the atypical 50V tap, then the 100K in front of the bias rectifier diode is moot. your B- will be so low, you'll cook the output tubes and possibly other things. 100R-1k is more suitable value for that part. 

I'm thinking more like 3K3 given 6V6's at 408V.

[Patrik]

And what about putting Fx loop one stage before reverb?

[2deaf]

And what about putting Fx loop one stage before reverb?

It's up to you, but most people seem to put the effects loop right before the power amp.

[Patrik]

So the problem you were talking about is output impedance? I tought it was input impedance.

[terminalgs]

extra triodes can be the bane of a nice design just like extra knob holes when you start trying to find a use for them.


you could use the extra triode as a mixer for dry and reverb wet, or put it in signal path between reverb  send-to and return-from so you won't don't need big 3.3M to defeat feedback  but .. you have mix wet and dry somewhere (and maybe an original Fender design is best?   all of these things will just complicate the circuit and probably lead to more issues to sort out and possibly sound worse.


If you must use it, consider parallel triodes as first gain stage,  or a CF follower after the gain stage like a bassman 5F6.



also, be sure to put a big grid leak resistor on the triode after the FX loop, so you have a big input impedance for whatever you have in the loop.  if you plan on long patch cords between amp and FX's,  you might need a grid stopper on the same triode. 

[Patrik]

Thanks for idea terminalgs, I completely forgot about the parallel input stage option and a sweet drive it can provide (talking from experiences with a Bassman RI), so one excess triode is used. For reverb mixing I was thinking about using an original 3.3M resistor value but without a bright bypass cap. As 2def explained previously FX loop performs better closer to power amp section so I was thinking about putting it between triodes of tube #4, (what would be third gain stage and phase inverter) and use 1M volume control potentiometer as a grid leak resistor. About the leads, they would be 1m (about 3ft) long in each direction at most as I just need them to connect to a reel-to-reel recorder I am using for echo/relay/reverb effect so no need for grid stopping resistors I think. What do you think of that configuration. Thanks.

[2deaf]

So the problem you were talking about is output impedance? I tought it was input impedance.

The amplifier Effects Send has an output impedance associated with it.  The effect device has an input impedance associated with it.  These two impedances form a voltage divider, but that is usually not a problem because the effect device input impedance is frequently much larger than the Effects Send output impedance.

The cable between the amplifier Effects Send and the effect device input has a capacitance and this is what creates a problem.  The longer the cable, the greater the capacitance.  This capacitance is in parallel with the effect device input impedance making the overall input impedance dependent on frequency.  This means that the voltage divider formed by the output and input impedances changes with frequency.  We don't want that.

The solution is to make the amplifier Effects Send output impedance so small that the voltage divider doesn't change significantly across the frequency range that we are interested in.

[Patrik]

Thanks, but how can I lower impedance?

[terminalgs]

you could put a 1M pot between the send/return and grid.  When effects loop not in use, you'll effectively have a master-volume. nd add a lot of complexity and if yo


also consider the input sensitivity of your tape machine.  you might overdrive it's preamp with a big fat post-preamp signal en route to the phase inverter.  your guitar output might somewhere between 10 and 100mV which can work with a 50s/60s tape machine usually fed a high impedance mike,  but if you give it 5-20V,  it might distort heavily.    Any reason not to let the tape machine sit between guitar and amp (and ditch the FX loop)?


 to me FX loops are always troublesome because if you design it for any random set of pedals, you have to include buffers and make up gain stages and the like, and  further open the door to dealing with noise, extra components and all the stuff that suck the life out of an amp.

[2deaf]

...FX loop performs better closer to power amp section so I was thinking about putting it between triodes of tube #4, (what would be third gain stage and phase inverter) and use 1M volume control potentiometer as a grid leak resistor.

If you mean the FX loop is directly feeding the cathodyne phase inverter, that won't work.  The FX Return needs to be amplified by a gain stage which then feeds the cathodyne because effects level signals are too small for a phase inverter that has no positive gain.   

[Patrik]

Thank you all for help. I did some thinking and realised that I was complicating stuff to myself for no reason. I had been using a R2R based delay for last three years feeding the signal directly in the front of the amp and have never encountered any issues. Therefore I decided to abandon the idea of Fx loop as I use no effects other than delay. So cancelling Fx loop from signal-chain should simplify the build a lot. I will add a photo of diagram bellow as soon as I finish drawing it. Thank you all.
   

[Patrik]

Here is the final (hopefully) diagram.

[sluckey]

Better put a resistor in series with those LEDs.

[terminalgs]

check out the schematic for the matchless chieftain for parallel triodes. you can do it without 220K mix resistors and a single plate resistor, single coupling cap and single  cathode resistor and bypass cap.  if not to just save component count, also your 2 mix resistors will lower gain when followed by the tone stack.

[Patrik]

Thanks terminalgs.

[Bieworm]

The dwell control I would add anyway. It replaces your 1M to ground at the input side of the reverb grid.

[Patrik]

Here is a modified first gain stage. It has lost mixing resistors and coupling caps altogether as I decided to use tone stack as a coupling cap. And thanks for suggestion to add dwell control.

[sluckey]

There will be little difference between that parallel triode and a single triode. IMO, a good use of an extra triode would be a good tremolo circuit. I know that many people don't care for tremolo. But this page may be interesting to you...

     http://sluckeyamps.com/trem/trem.htm

[Patrik]

I know that tremolo is a nice adition to the soun if used properly, but I just don't use it for playing and recording.

[sluckey]

GZ34 was replaced with full bridge rectifier.

You can't do that. The left two diodes will pop as soon as you turn on the power, possibly even destroying the PT.

And as previously said, you must have a resistor in series with those LEDs. Otherwise, they will blow out immediately.

[sluckey]

...

[Patrik]

Here is digitalised diagram with GZ34 replaced with diodes, but for build GZ34 wil be used, was omitted due to lack of sutable symbol in software. Hopefully in about two months (at most) Iwill be able to post a photo of finished amp. Thanks for all the help and advise guys.

[thetragichero]

I'm a big fan of using 12dw7 for reverb (i use the 12ax7 triode for reverb recovery)

[2deaf]

DC heaters aren't really necessary for this amp.  If you are going to use DC heaters, only V1 and possibly V2 would use the DC while the rest of the tubes would use AC directly from the winding.

An artificial center tap for the winding would be beneficial.  Don't ground the DC portion.

The large current draw on the DC portion necessitates large filter capacitors like 10,000uF.

I always avoid diodes of any kind in parallel because one of them may take down the lion's share of current while the other starves.  I recommend a separate 270 ohm resistor in series with each LED as opposed to one resistor for both.

[Patrik]

I won't be using DC heaters, I didn't include them on diagram as I didn't find symbol in the software I used, I replaced GZ34 with diodes for the same reason. Caps in DC setup are used just to reduce/eliminate LED flickering as they will be used as a backlight for control pannel and I just wrote in a value I liked. For build itswlf I willjust use some salvaged caps from my heap of old electronics.

[2deaf]

I won't be using DC heaters, I didn't include them on diagram as I didn't find symbol in the software I used, I replaced GZ34 with diodes for the same reason. Caps in DC setup are used just to reduce/eliminate LED flickering as they will be used as a backlight for control pannel and I just wrote in a value I liked. For build itswlf I willjust use some salvaged caps from my heap of old electronics.

Humans can't see 50 flickers per second, so there is no need for a capacitor.



 

[PRR]

> Humans can't see 50 flickers per second

Standing still (light AND eye), no not really.

Moving, I sure can see 60 and even 120Hz flicker. I knew a conductor who objected strongly to all fluorescent lighting, because his baton looked janky.

[Patrik]

Yes, humans ordinarily can't see flickering at 50/60Hz but cameras can and this is my main concern as I do a reasonable amount of filmed live performances.

[thetragichero]

if people are focusing on a 5mm led at distance you might need to focus more on practice and or stage presence than an indicator light?

[Patrik]

These lights are more of a help for me than an audiences object of interest. And I am planing on making something similar to Echolette Showstar amps from the 1960s from a control panel perspective.

[sluckey]

if people are focusing on a 5mm led at distance you might need to focus more on practice and or stage presence than an indicator light?

That wasn't very nice. Certainly not helpful.