This is an interresting datasheet for your project with a 3,5K transformer. If you can get 375V from your supply, you can have 35W with a cathode bias and very low distortion. I notice they use a separate 260 ohms cathode resistor for each tube, maybe, that's why I don't get a nice waveshape in my spice simulation because I use a common resistor for both tubes ...
I will try it with those Mullard specs to see the difference and also show you my phase splitter idea when I will not be too busy ... I don't finish my outside work yet and I also have to repair my 1990 chevy motor before winter ... Construction and car mechanic are other kinds of DIY I like to do !
Alain.
Huh! Chevy's are really a big problem! (ha, that was a joke!) I'll wait for your PS design, although I already have mine (cathodyne type, etc). I'll probably try both, cathoded and fixed bias.
Yeah, I find with guitar amps it is often times better to start with a design that sounds good and then tweak it from there. Many of their "larger" amps are dual EL34 designs, so you can browse their site, listen to some of the demos on each, see what you like, and then Google a schematic. It will save you a lot of time and you will know that it will at least have good tone when you first use it.
My only tips are to use tube rectification with an input cap no larger than the data sheet recommendation, and use an output transformer a tad on the smallish side (i.e., 30 watts for a 35 watt amp). Power supply sag and output saturation effects are key parts of a guitar amps tonal character.
That's strange, I thought on using larger OT, don't want core saturation. Once I repaired a Bassman 135, that had a 50W OT, then changed it to a 150W OT and started to sound really good. Smaller OT sounded really bad, and low frecuencies sounded really distorted.
The idea of 30-35W is because I can push it and loudness could be tolerated. My SS FET based guitar amp has 100W output... and I can crank it, but then I can't listen what I think! BTW, I'm curious about how could sag sound in a SS amp, or how can I make it (maybe with a smaller power transformer?) but that's other story...
We were playing the guitar in that Bassman amp, that has UL OT, and low frecuencies were not really dynamic. I'm looking for versatility, depending on what my preamp will deliver, but really want from warm clean, to mid or heavier crunch. EL34's really help with all this I want. I have to build it first, and then we'll see.
Well, like I said there is no hard and fast rule for guitar amps. But the fact is that the combination of the amp, OT, and speaker all make up the character of a guitar amp. You may end up finding yourself experimenting with a few different transformers, especially if you get them surplus (or otherwise cheap).
Yes, all guitar amps are made for specific kind of music : Classic, rock, metal, etc.
Usually, a saturation effect is obtain with a tube preamp clipping, the 12AX7 is prefered for this task. This very good paper explain how it work and so much more :
http://www.freewebs.com/valvewizard1/Common_Gain_Stage.pdf
I believe all tubes preamp designers should read that, specially for guitar amps ...
Happy Halloween !
Alain
Last edited:
Next time, I will buy a Toyota ... The parts cost twice but they last 10 times longer ... The only good things about my old chevy are : It is build like a tank and I never learn about car mechanic so fast trying to repair it myself.
Happy Halloween !
Alain.
I try to stay away from the 12ax7, although this preamp will use it, also gonna try some at7's. I wish I could get some surplus OT, they're not quite cheap around here!
You should try a E30 BMW, they are really well constructed cars, and they provide a sport driving. There's a 4WD version of one of them. Bad point is they're from the mid to late 80's, but sure they are cheap up there.
You should try a E30 BMW, they are really well constructed cars, and they provide a sport driving. There's a 4WD version of one of them. Bad point is they're from the mid to late 80's, but sure they are cheap up there.
A BMW ... Noooooo ... I am not rich enough to buy a car like that, my budget for a new car is around 1000$ and I try spending less than 300$ a year for repairs and maintenance ... Having a car getting very expensive.
In Quebec, the gaz cost about 1,45$ for a liter this month, those petroleum company are real rip off ... I hope the "Occupy" movement all over the world in 1500 big towns will change something one day but I am effraid it gona take much more than those pacific "sit in" to stop the terrible voracity of the 1% who own 80% of the wealth of our planet ...
Alain.
I made a stupid error on this schematic ...
The plate dissipation at full power is wrong because it include the output power and power loss in the primary of the output transformer ...
Here's the corrected schematic :
I measure carefully the mean voltages and currents with the simulator, I use mean and not RMS because the waveforms are not very sinusoidal, a DC voltmeter and ammeter should give about the same results.
As you notice, the plate dissipation for each tube is now only 8,37W and not 24,568W ... The current measure where is IPROBE1 (R9 on the other schematic) is 143,5ma, since there is only 342,7V between plate and cathode, the total power is 49,1W ... But if you substract the 29,8W output power and 2,6W winding loss and divide by two, you get 8,37W ...
The plate dissipation at full power is higher than with no signal but it is not always like that, sometime, it is lower ... It depend of the class and operating point ...
So, you have to place the DC ammeter where is R6 or R7 to measure the current and multiply by the voltage between plate and cathode to get the plate dissipation.
Alain.
An externally hosted image should be here but it was not working when we last tested it.
The plate dissipation at full power is wrong because it include the output power and power loss in the primary of the output transformer ...
Here's the corrected schematic :
An externally hosted image should be here but it was not working when we last tested it.
I measure carefully the mean voltages and currents with the simulator, I use mean and not RMS because the waveforms are not very sinusoidal, a DC voltmeter and ammeter should give about the same results.
As you notice, the plate dissipation for each tube is now only 8,37W and not 24,568W ... The current measure where is IPROBE1 (R9 on the other schematic) is 143,5ma, since there is only 342,7V between plate and cathode, the total power is 49,1W ... But if you substract the 29,8W output power and 2,6W winding loss and divide by two, you get 8,37W ...
The plate dissipation at full power is higher than with no signal but it is not always like that, sometime, it is lower ... It depend of the class and operating point ...
So, you have to place the DC ammeter where is R6 or R7 to measure the current and multiply by the voltage between plate and cathode to get the plate dissipation.
Alain.
Thanks again! Could you try it with 370 or 380 VDC? I'll be almost the same. I'm really busy this days, but can't wait to build it! Really! I've been searching for real primary impedance, I've seen it could be 6k, or 6.6k... don't know exactly, really... Remember Mullard's datasheet says 3.5k!
Well, I'm very busy too for many more weeks, but I'm not sure you can get 370 to 380V with your supply, there is a big voltage drop when there is a big load like in your amplifier project. You should try it first with a 250ma load using power resistors. If you think you can get about 380VDC at 250ma, you need a 1500 ohms resistor ( 380 / 0,25 = 1520 ), the dissipation will be 95 watts ( 380 * 0,25 = 95 ) then you should use a power resistor of over 100W rating.
You can combine many power resistors to get this value and power rating, for example, three 500 ohms 35 watts (or more) resistors in series will be OK. I always do that myself before using any transformers in my projects, sometime, I have to use over 10 big resistors in series/parrallel combination the verify the output voltage of my power supplys and/or power transformers with a particular load. If you use different values and power ratings resistors, you have to calculate the power for each or some will stay cold and other will burn ... Ohm law : E = R x I and P = E x I ...
If you like to know the real impedance of your transformer, don't search, don't assume and don't guess, measure it yourself, it is very easy ...
Just hookup any low voltage transformer to the primary of your OPT with no load and measure the voltage at is primary and 8 ohms secondary, divide the primary voltage by the secondary voltage of the OPT and you will get the voltage ratio ... The impedance ratio is the square of the impedance ratio.
Example : You read 10VAC across the primary and 0,4VAC at the secondary, that is a 25:1 voltage ratio ( 10 / 0,4 = 25 ), then the square of 25 is 625 ( 25 x 25 = 625 ), this is the impedance ratio, if you multiply by 8 ohms you get 5000 ohms ( 8 x 625 = 5000 ) ...
You must verify those two things before doing anything else, it is the first step !
You cannot design and build an amplifier without measuring voltages and currents and doing some maths ... It is impossible unless you use an already existing circuit and the exacts power and output transformers brand and model needed ...
Alain.
It is a good thing you start to build it, this is the bigger part of the job. Adjusting some parts values to get the right voltages, currents and dissipations are details. I think you can use your actual 5K OPT with no problems, there is not so much differences with a 3,5K like in the Mullard datasheet.
I made four simulations to show the differences using a 5K OPT instead of a 3,5K OPT, with fixed bias and cathode bias ... I use the typical applications values from the Mullard datasheet, the currents are different because my 6CA7 Spice model is not perfect but this is also true for the real tubes, they are never exactly like the datasheets said.
Notice the maximum plate dissipation is 25W, the maximum screen grid dissipation is 8W and the maximum cathode current is 150ma ...
An externally hosted image should be here but it was not working when we last tested it.
With a average quality 3,5K OPT and fixed bias, the output power is 39,75W ... The Mullard datasheet said 48W but that's with a perfect OPT and identical tubes caracteristics ...
The only problem is the 8,89W screen grid dissipation exceed the 8W maximum but that's only at maximum power with a steady sinewave signal ... Then it is acceptable but in real life, a higher value than 600 ohms for R1 will correct this problem.
An externally hosted image should be here but it was not working when we last tested it.
With your 5K OPT, the power is a bit higher but the screen grid dissipation is much higher than with the 3,5K OPT ...
An externally hosted image should be here but it was not working when we last tested it.
With the 3,5K Mullard OPT and cathode bias using separate 260 ohms resistors, the output power is 31,63W, the Mullard datasheet said 35W ... The 7,432W screen grid dissipation is lower than the 8W maximum at full power but the 28,9W plate dissipation with no signal is much higher than the 25W maximum and that is very bad for the tubes ...
The only way to make this dissipation lower is using a higher bias voltage, then using a higher value for the R2 and R5 260 ohms resistors ...
An externally hosted image should be here but it was not working when we last tested it.
With your 5K OPT and cathode bias, the 35,1W output power is about the same as Mullard said but the plate dissipation with no signal is too high like in the preceding schematic and the 9,332W screen grid dissipation is over the maximum too ...
So there is a double problem with this one and they can be corrected by using a higher value for the screen grid 470 ohms voltage dropping resistor and the 260 ohms cathode bias resistors.
Like you can see, you don't have to worry about the screen grid dissipation with no signal and the plate dissipation at full power, thoses are always much lower than the maximum ... My simulator show a lower distorsion than 1% for all four circuits.
You will have to adjust the currents of your circuit anyway because datasheet curves and Spice models are just for general design guideline and I am sure you will have a much lower supply than 375V because you measure it with almost no load ... I think you will get only between 300V and 325V ...
Alain.
Thanks again! Actually I'll use 1k grid stoppers, and I'll use fixed bias. It's curious how this tubes behaviour changes with minor changes. It depends on what tubes I'll use, too, I watched a EL34 based guitar amp with different brand EL34's, and output power changed a lot when tubes were changed.
Only thing left is building it!
Only thing left is building it!
Different brands tubes need different bias to get the same currents, operation point and output power. Same thing for same brand tubes but different matching ranges called "soft", "medium" and "hard" ... The best is using a different bias adjustment for each tube because they are never perfectly matched and OPT are often unbalanced causing amplitude distorsions.
Alain.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.