Hi All,
I am learning stuff everyday on tubes and it's a blast! I am still learning how to use the oscilloscope, so, I can't say for sure if my design is as clean as I want, but, it sounds very good to my ears.
Here is the link to the lastest design (Sopht 12K5 - slight return part 2):
http://www.sopht.ca/index.php?s=26
And here is my first "theory" document. All comments / corrections are greatly appreciated.
http://www.sopht.ca/index.php?s=35
Thanks,
Stephane.
Wow, very cool! the theory bit is great! i was just wondering, what kind of sound you get from the updated 12k5 design if you stick a treble booster before it?
BTW, have you built a finished sopht amp? Coz if so, can we see some pics :P
Are you going to do a theory page for the 12k5? :P
Hi Alpha,
The sound right now is very clean and very articulate! I dontt think that at a gain of 20 (max), I will be able to overdrive the power tube, so, I am going for a very clean sound. With a clean sound, it's easier to put a nice distortion pedal infront of the amp. Actually, when I tried the sopht amp with a bk butler tube overdrive, it was awesome! I don't have a trebble booster and therefore, I can't try that setup. I am planning on adding a simple tone stack (BMP style tone stack).
Unfortunately, I am still at the design stage and therefore, I don't have a complete build yet. I can post pictures of my setup though :).
I will do a theory section about the 12k5 as soon as I sort it all out.
Cheers,
Stephane.
Very cool--a tube miniamp is now definitely on my to do list.
Some questions: How is current draw--do you think it would be possible to actually run this on batteries, say either a small lantern battery or a rechargeable power tool power pack with a step-up, for a reasonable time per battery/charge?
Also, have you given any thought to either a push-pull output or a pentode input stage?
Anyway, time to start keeping an eye out for a suitable enclosure...
Ben
How do you use the tube chart to figure gain? I've got myself 2 12U7's and a 12SK5 as well. I've been thinking about an a/b push pull design for use with two 12SK5's but want to start small first. I prefer a little bit of distortion in my amps as opposed to clean (actually, alot of distortion) and want to know how to tweak the 12U7's for my desired results.
I have prototyped slajuenes latest version of the 12U7 circuit but only the dual triode preamp section.
I haven't actually listened to it yet because I got hold of a frequency generator progam and an oscilloscope program for the PC today so I should I'd try it out first.
When I view a sine wave directly (at any frequency) from the generator and compare it with the output from the preamp on the dual scope setting.
I can see that the preamp is nicely rounding off the peaks of the sine wave but only on the top of the waveform the bottom still looks the same
- Am I correct in that both tops and bottom of the wave forms should be smoothed.
BTW the smoothed top peaks become progressivly more squarish as I increase the output level of the sig-gen.
Any idea what is wrong here? I will continue debugging tomorrow evening and give the pre a listen as well but if any body can give me a clue I'd be grateful as I am new to this.
It's great being able to see stuff with the scope - even if it is only software and works through the PC soundcard.
/Taz[/url]
BTW : the oscillocope software and signal generator is freeware u can get it at
www.dazyweblabs.com
loads of other freeware stuff there - sorry if this link has been posted before.
/Taz
Hi All,
Ben, the major current draw is from the heaters (.15 for the 12U7 and .45 for the 12K5). That amounts to .6A just for the heaters. The rest is pretty negligeable. A 1A wallwart is more than enough. Actually, we could probably get by by using a 800ma. One could use a lead acid battery. Some are rated pretty high (i.e. close to 3A!!!).
I actually have thought about a push pull using 12K5. I got a beefy power supply for that (12V 2.4A). Because of it being a push pull, the minimum current draw is 2X.15 + 2x.45 = 1.2A. But, I haven't finnished my single ended explorations yet. Might actually go back to the 12CX6 before moving on to a push pull design. I have thought about using either a 12K5 or a 12CX6 as a preamp. Although, I would probably try to put it after a 12U7 to try and push the 12K5 harder.
William, you don't use the charts to calculate gain. You use several math equations + specs from the spec sheet. Here is a nice link to calculate the gain: http://www.rru.com/~meo/Guitar/Tubes/t2.pl . If you are using a bypassed cathode capacitor, the cathode resistance is considered null. For example, a 12U7 with a cathode bypass cap with a plate load resistor of 100k, a plate resistance of 12500 and the transconductance of roughly 1600 would give a gain of 17.6. Altough, I'm not sure that I am actually getting a gain of 17.6.....
The only distortion I can see you generating with one 12U7 and one 12K5 is preamp distortion. You can bias the 12U7 to be closer to saturation (i.e. grid closer to 0v, say -0.25v). To do this, simply decrease the cathode resistor until the grid voltage is close to -0.25v (or higher, the higher you go, the more distortion you'll get). Let me know how it works out!
Cheers,
Steph.
Yea, I'm actually planning on using two 12U7's. That should give me 4 gain stages, and if I still don't get enough, I'll resort to diodes. But I'd prefer using just the tubes. Thanks for the link, I'll check it out right now.
Hi Taz,
cool, your one step ahead of me (I'm planning on putting my scope to work this week!!!). Basically, what you are seeing is the tube clipping the signal. What you are seeing is the tube reach saturation (top part of the sine wave). Therefore, this setup would generate preamp distortion. BTW, what is your output level (i.e. 100mv or bigger?!?!). A typical guitar generates a 100mv signal.
Thanks!
Stephane.
Hi All,
I had a bit of time to play around with the amp again tonight. My setup stayed pretty much the same as what I have posted as my slight return part 2 but I added a second bypass cap on the cathode of the second triode. This produced a nice smooth distortion. Here is a link to the clip:
http://www.sopht.ca/sounds/sopht_12k5_v3.mp3
Ben, I know that you are looking for WAY much distortion than this, but, it's a start. If you look at really big gain monsters (soldanos, boogies, bogners, etc), they tend to generate a lot of preamp distortion, so, with 2 x 12U7, you could make a serious attempt a that! I might try something similar, but for now, this sounds nice to me!
Enjoy,
Stephane.
Slajeune: I can vary the output as my signal generator is PC software so I can vary thefreq from the program and the output level from the PC soundcard volume control.
So by starting at very low outout vol (from the PC) and increasing it I can see the tube rounding off the sinewave and then going into clipping as I continue to increase the output level but I still don't understand why only the top part of the waveform (the positive going part) is clipped and not the tops of the negative half of the waveform - have I biased the tube incorrectly?
/Taz
Hi Taz,
Well, in typical class A setup, both parts of the sine wave should be cutoff (i.e. top and bottom should be cutoff). So, in that regards, I think that yes, the stage seems to be mis-biased. If anybody has a better explanation (or if I am mistaken), then, please jump in!!!
BTW, what are the readings for the grid voltage for that triode stage? Calculate the grid voltage by putting the negative probe on the cathode and the positive probe on the grid (for stage 1, cathode = pin 3 and grid = pin 2).
Cheers,
Stephane.
triodes clip for 2 reasons:
1.: when grid is very negative, tube cuts off = no conduction = plate voltage stays at max. value, can`t get higher;
2.: when grid is zero, tube is at max. conduction = plate voltage is at min. value, can`t get lower;
if biased right in the middle of these grid voltages, "clipping" occurs at both halves - however a little different on both sides.
(roughly said...)
:idea:
that sound clip sounds great!
stephane and puretube - thanks - I'll try out your suggestions when I get home after work.
Alpha - i don't agree with you - i don't think the sound clip sounds great - .........it sounds absolutely fantastic.
Quote from: tazwolf.........it sounds absolutely fantastic.
I'll second that! :D 8)
Take care,
-Peter
That sounds great!
Steph, I actually wasn't looking so much for a lot of distortion (that may have been someone else), just a way to get power tube saturation--that is why I suggested a pentode preamp. I guess I'm thinking along the lines of Dr. Z amps that make that gorgeous sound with next to no preamp distortion at all, like the Rt. 66 and Z-28. It's all about getting big clean gain at the input, so maybe a cascode would also do it. But what you have there does sound terrific.
Ben
Hi Ben,
you're right, William was looking for max gain. I am also interested in power tube saturation as oposed to preamp distortion. I will have some time this week to hookup my oscilloscope to it to do a bit more debugging.
Cheers,
Stephane.
Stephane
My grid voltage on the triode section is -0.45 on the second one it is -0.54.
I experimented by removing the the 0.068uf cathode bypass capacitor on the second triode and it got rid of the uneven clipping (clipping only the positive going part of the waveform) that I was talking about in my previous posts. This however results in a lower volume. It also brings the grid voltage on the second triode to the same as the first -0.45 volts.
So now I'm waiting for my 12k5 tubes to arrive so I can actually listen to it thru a speaker - I might end up adding the bypass capacitor via a switch and experimenting with a few different values. What exactly does the this capacitor do?
I was reading thru the sopht theory and I think I understand about biasing for class A operation (and so the grid voltage should be half way between 0 and -1). My question is if one had a higher voltage on the plate then the voltage swing would be higher - according to the chart for the 12U7 (if I read it correctly) at 12v on the plate the swing would be between 0 and -1.5 thus resulting in a grid voltage of -0.7. Would this be better in terms of output level? or have I completely misunderstood.
BTW The guitar going through just the two 12U7's sounds pretty nice.
I just read up on cathode bypass resistors again for the 3rd and I think I finally understand what they do - stabalise the grid voltage!
Hi Tazwolf,
First, the cathode resistor bypass cap. You are right, it is used to stabilize the voltage on the cathode. While doing this, it also increases the gain considerably (roughly 2x!). That explains why without this cap, the volume is lower. To see it's effect on gain, go back to the link I gave to calculate the gain of a triode stage. In the cathode resistor entry, put the value of the cathode resistor if you aren't using a bypass cap or put 0 if you are using a bypass cap. You can see it's effect on gain.
As for the question on plate voltage, I have "two" conflicting versions. One says that we should use the voltage coming from the power supply as the "plate voltage" on the chart (in our case 12v) and one says that we should use the voltage after the plate resistor (in our case, the plate resistor is 100k) which would make our voltage 4.8v.....
The difference between both is the bias point. If we take 4.8v as our plate voltage, the bias point is -0.5 which gives us a 7k cathode resistor. If we take 12v as our plate voltage, the bias point would be -0.75 which would give us a cathode resistor of 12k.
For what it's worth, the latest sound clip I posted in this thread had a cathode resistor of 12.2k on the first triode and 14.7k on the second triode. This would lead me to think that the second method (i.e. 12V on the plate) would seem to be the most appropriate way for our needs. I still need to see the output on the oscilloscope though...
As for using a smaller plate resistor, everything that I have read in regards to space charged tubes points to using large plate resistor values. Look at the data sheet for the 12AL8, it shows a 330k plate resistor to it's triode stage!!!!
One of my plans is to make the second bypass cap "swithcable". Therefore, I can go from clean to dirty with a simple flick of a switch!
Puretube (or anybody that has lots of tube experience), is any of this making sense?
Thanks,
Stephane.
that switchable cathode-cap is exactly what a lot of old (and new) prof. amps got, named "Boost" . Often a medium value is chosen, to boost only the mids/highs. Sometimes a clever R/C combination for just mid-boost.
Stephane:
So using 12v at the plate and biasing would maximise the potential of the tube - in theory. Did you have any problems with ths - if I just remove the 100k resistors with jumpers and use trimpots for the cathode resistors I can't get the grid voltage bias to anywhere near -0.75 using the values u mentioned 12k and 14k. I get a value of 5.2k on the trimmers with -0.75 as the grid bias voltage and the tube just seems to refuse to conduct. - but I'll keep working on it.
All this is rather addictive - I have to remind myself to go to sleep cos I have to work the next morning or to take my motorbike out for a spin. Soundwise I think you have really come across a great idea with these 12v tubes Stephane even though I haven't come to the setup I am satisfied with yet I have not been this inspired by an amp sound since I played a gig with my SG and an ac-30 on full whack...but that was a long time ago.
Puretube: And there was me thinking it was another preamp stage on those amps.
/Taz - another beer another grid voltage bias
QuotePuretube: And there was me thinking it was another preamp stage on those amps.
he he, that`s the 2nd (or "hot")
channel :)
which one usually ends up paying "hot" money for as it's an extra feature over the standard model!
/Taz
Hi Tazwolf,
ok, sorry about the confusing post. the 12 and 14k are both cathode resistors. Don't remove the plate resistor (i.e. don't jumper the 100k resistor). The plate resistor needs to stay at 100k, only play with the cathode resistor (for now). You bias the tube with this resistor. So, I would use trimpots (or simple pots) for the cathode resistors. Again, I'm sorry I haven't had the time to try any of this oscilloscope thing yet. Too many things, too little time.....
Thanks for all your help Taz!
Stephane.
slajeune, at the moment, how is the 12k5 biased (ie some voltage/current readings)?
BTW, on your theory page, how did you calculate the voltage drop across the 100k resistor?
Another quick question :) ,
What would it sound like if you took the schematic of a Marshall plexi, and replaced all preamp tubes with 12u7's, and power tubes with 12k5's, and rebiased to work on 12v/24v?
crap... (sorry)
lol, just a thought :P
I'm getting slightly fustrated by my inablility to comprehend how to read tube charts. I'm trying to adapt the gain stage of the 12AX7 using the 12U7's that I have. I realize the 12AX7 is a high MU tube and the 12U7 is a med MU tube, but the desire still resides.
The gain stage is set-up as,
(http://209.240.71.22/circuit/slo10012ax7.gif)
and is the typical configuration for Soldano Super Lead Overdrive amps, such as the actual Slo100, Mesa Boogie Dual Rectifiers, and Peavey 5150's.
How do I go about reading the charts for the 12AX7 and 12U7 to see how they're biased and get close to the results I'm looking for. I've re-read the book I have on tube amp design, but the author used a butt load of math without explaining where he was getting the values or why.
I guess this also goes into the territory of getting a plexi/fender/matchless/Vox/insert amp here type sound as well......
check here:
http://www.glass-ware.com/tubecad/,
and read thecomplete http://www.tubecad.com
lots to learn! (although maybe a little advanced, sometimes....)
Hi Alpha,
concerning the 12K5, I will get back to you, I haven't checked the voltages recently.
I didn't calculate the 100k voltage drop, I observed it with a DMM. I whish I could calculate it though...
I will update the theory page. I read a second article that confirmed that we should use the voltage comming from the power supply as our "plate voltage" and not the voltage after the plate resistor. Therefore, the cathode resistor should be roughly 12k. Again, this is what I currently have (ok, 12.2k and 14.7k....) and this is what was used to record the last clip.
As far as voicing is concerned, from my understanding, a good part of the tone is molded in the preamp. Therefore, great care must be taken to "replicate" the preamp section of your favorite amp. Having said this, the 12U7 is a different tube than say a 12AX7 or a 12AU7 and therefore, will sound different. Also, the gain is a lot lower (20 vs 100 for the 12AX7). Since we seem to be using larger cathode resistors, the bypass cap will need to be adjusted in order to achieve the proper voicing. From what I gather, the cathode resistor and the bypass cap, form a filter which modifies the voicing of the amp. This and the coupling caps would definately change the sound of the amp. One needs to play around with them to find the right tone for them.
Cheers,
Stephane.
just on an off note, does anyone kmow where i can get a power transformer that has a variable primary for 112v/220v/ etc, and has a 24v/0v/24v @3A and a 5v @ (howerver many amps a valve rectifier sucks) secondary?
I been reading the tubecad page as suggested by puretube - it goes through triode biasing without a plate resistor. If one tries the formula for biasing with values for the 12U7 then it cmes out that it is not possible to do so (the cathode resistor would have to be negative) with a B+ of 12v.
If you try it with a B+ of 12.5 thewn it works (I haven't got a power supply greate than 12v so I can't try it in reality).
The other thing that struck me is that the plate resistance (resistance of the actual plate which is 12500 ohms according the datasheet) is used in the formula for calculating the cathode bias resistor when there is no separate plate resistor - but why is the internal plate resistance not used in the other formulas for calculating bias?
/Taz
I've been thinking.... :shock: about the bias and setup of the first triode.
In the experiments I've been doing - the grid bias has been set in order to give the maximal output swing - in this case - 0.75 without plate resistor or -0.5 with plate resistor. If I am using a guitar as an input then am I really gaining (no pun intented) anything from having the bias at such levels as a guitar input will not be able to swing the grid anywhere near + or - 0.5volts.
So is it better to bias the grid appropriately for the input coming into it? in this case bias the triode so that approx plus or minus 0.05v going into the grid will give maximal output swing. As i have it setup at the moment the first triode is givng me 4x gain from a guitar level signal with no distortion (as seen on a scope using a sine wave input) with the grid bias voltage set to -0.1.
Sorry for all the silly questions but sitting and messing with 50 year old tubes is just so much fun that it would be a shame not to try and understand how they work.
/Taz
soz, but what is the average output pk2pk from a guitar?
Alpha: slajeune posted it in this thread a way back - 100 mv or there abouts.
/taz
Hi Taz,
concerning the bias of the first gain stage. I guess you could bias it as close to 0 as you can (i.e. -0.05). The only problem I see with this is you don't have enough 'headroom'. Say you want to use a booster or a pedal that boosts your guitar, the first stage will automatically go into asymetric cliping. If it were properly biased (i.e. mid point) it would be able to cleanly handle the boosted input and if it did go into cliping, it would be symetrical. I don't think that having the tube closer to 0 change it's gain, only it's operating point.
Thanks,
Stephane.
One thing, isnt it good to have some asymetrical clipping to get even order harmonics?