One thing to keep in mind...
Audio power amps and RF power amps are two totally different mind sets for designing.....
With RF frequencies you use SMITH CHART and 'usually' reference everything to 50 ohms......and you look at LAMBDA and it's fractions..for example 1/4 lamba....with respect to transmission line reflections...whether thier cables, micro-strip or waveguides...
With RF amps you use complex conjugate matching on both the input and output ports....and you design very a narrow bandwidth, usually one frequency.... You are trying to achieve maximum power transfer and at high RF frequencies impedances become unusable...so you use S-PARAMATERS... S21 (forward transfer Gain) for example...ect...ect...
There are cases in RF where you can have a single ended Class AB circuit.... This means the device is biased in AB, close to cut-off, but when RF power to driven into the Gate or grid....RF energy can rectify and re-bias into Class A when it is driven ....
Audio amps ...when setting the plate load, you are trying to get the 'optimum plate load' ....this means best power output for lowest distortion.....
The old rule of thumb is plate resistance x 4 for TRIODE plate loads....but that does not always produce the best optimum design...
Your tube is a 0-bias transmitting tube....If you look at the curves for this tube, they do not fit the norm TRIODE curves as commonly used in audio amps.... These curves almost look like Ultra-linear curves....so you need to set the plate load by examining the plate curves and fitting various plate loads to find one that provides the best compromise of power output, bandwidth and safe plate dissipations... This tube was designed for Class B, C operation...this provides a lot more time for the plate to cool durring the longer cut-off time.... When using this in Class A , you need to re-consider maximum dynamic plate dissipation...
What do you plan on driving the grid with????? You would need a 6550 cathode follower just to over-come the Miller capacitance for large signal drive
Chris
Audio power amps and RF power amps are two totally different mind sets for designing.....
With RF frequencies you use SMITH CHART and 'usually' reference everything to 50 ohms......and you look at LAMBDA and it's fractions..for example 1/4 lamba....with respect to transmission line reflections...whether thier cables, micro-strip or waveguides...
With RF amps you use complex conjugate matching on both the input and output ports....and you design very a narrow bandwidth, usually one frequency.... You are trying to achieve maximum power transfer and at high RF frequencies impedances become unusable...so you use S-PARAMATERS... S21 (forward transfer Gain) for example...ect...ect...
There are cases in RF where you can have a single ended Class AB circuit.... This means the device is biased in AB, close to cut-off, but when RF power to driven into the Gate or grid....RF energy can rectify and re-bias into Class A when it is driven ....
Audio amps ...when setting the plate load, you are trying to get the 'optimum plate load' ....this means best power output for lowest distortion.....
The old rule of thumb is plate resistance x 4 for TRIODE plate loads....but that does not always produce the best optimum design...
Your tube is a 0-bias transmitting tube....If you look at the curves for this tube, they do not fit the norm TRIODE curves as commonly used in audio amps.... These curves almost look like Ultra-linear curves....so you need to set the plate load by examining the plate curves and fitting various plate loads to find one that provides the best compromise of power output, bandwidth and safe plate dissipations... This tube was designed for Class B, C operation...this provides a lot more time for the plate to cool durring the longer cut-off time.... When using this in Class A , you need to re-consider maximum dynamic plate dissipation...
What do you plan on driving the grid with????? You would need a 6550 cathode follower just to over-come the Miller capacitance for large signal drive
Chris
nhuwar said:
The Cathode follower is non-phase inverting....buffer stage...
The Gain is less than unity, generaly in the .93 range, it can be easily calculated....
It is used when you need to have a low impedance source to drive a big signal voltage such as into the grid of a tube that has serious Miller capacitance...
The follower has some linearity issues...so to help fix that you can use a current source in the cathode.... Or you can do what McIntosh did...apply a unity signal positive feedback back into the cathode resistor to "bootstrap" the follower ....this makes the cathode resistor appear much bigger in AC impedance than it actually is, thus making it look like a current source.... Back then this was a great trick sinc adding another tube as a current source was not that cheap and easy..now aday you can simply throw a FET in there and be done with it...
First figure what your load will be..then you figure your voltage gain of the tube into you plate load...then calculate the Miller capacitance looking into the grid at full drive conditions...This will tell you what the input grid impedance will be at full drive and at your highest frequency of interest...no less than -3dB at 66kHz in mu opinion...shoot for 100kHz...
Once you know your worst case grid impedance and at what drive voltage level..you then design the driver stage to meet these requirements..
Chris
Since a single ended amplifer has one (or a few in parallel) tube, it must conduct some current for the entire audio cycle in order to remain linear. This is the definition of class A. You could build an A1 amp (no grid current) or an A2 amp (some grid current flows for all or part of the audio cycle). The 3-500Z is a "zero bias triode" which would want to operate in A2 for all but very high plate voltages. The efficiency of a class A amp runs from 5% to a theoretical maximum of 50% in A2. 8 to 20% is most common.
In a single ended design all of the output tube's current usually flows through the output transformer. This means the transformer must deal with a constant magnetic field caused by this DC current. This causes the core to operate at or near magnetic saturation, which will not be useful for linear coupling. To combat this the core contains an air gap and 5 to 10 times more iron. As transformers get larger it becomes harder to get a wide enough bandwidth for the entire audio range. The exception to this is a "parafeed" circuit which couples the DC into the output tube through a choke, and the audio is extracted through a capacitor into the output transformer (similar to an RF amp). A push pull transformer can be used in this application.
I have been told that the upper limit for a practical SE transformer is 50 to 75 watts. I had a custom transformer made for the 833A at the 75 watt level. It was not cheap. I tested it in a prototype design. It did not work well enough for HiFi, but it will make a killer guitar amp. It can easilly pass 200 watts over the guitar frequency range. The details can be found here:
http://www.tubelab.com/833SE.htm
There is a Yahoo group for people contemplating "extreme" amplifiers:
http://groups.yahoo.com/group/GM70/
You must register and be approved to read the forum. It takes a few days. I asked about BIG SE output transformers and got the same two answers that I have heard here. The Hammond 1642SE (75 watts) and the Electra-Print CU5KB (50 watts). The Hammond weighs 28 pounds, so we are talking about a large, heavy amp. I have not talked to anyone who has actually tried one in a high powered amp though, and some Hammond SE transformers have a less than stellar reputation. Electra-Print transformers are top quality, and they do custom transformers. They are not cheap. I asked about custom 100 watt SE transformers for the 833A and was told that a pair would be $1000, and that was almost 2 years ago. Transformers have gone up in price.
If you are looking to build a 100 to 200 watt amp I would consider a push pull design. The typical amp in this power range uses paralleled audio tubes like the 6550 or KT88. Four to six tubes per channel is common. This requires a low plate to plate load impedance and off the shelf transformers are available up to 280 watts. I have also used TV sweep tubes to make high power. 4 6LW6 (or similar sized) tubes per channel will make an easy 250 watts with the same transformers.
I understand the fascination with big transmitter tubes, so I would use a pair of 813's in push pull. These could do an easy 400 watts at high plate voltages, but again you are looking at a custom transformer due to the high impedances (20K ohms) and power levels involved. 2 or even 4 813's per channel at lowish (1KV) plate voltages and might work with available transformers. There are other possibilities like a pair of 211's, 811A's GM70's or similar tubes.
I am not sure what could be done with the 3-500Z's but a pair of them could provide enough power to set your speakers on fire if a proper transformer could be found. The data sheet shows 1420 watts in class B with an 8.6 K ohm transformer at 3KV. Running in class AB with a lower plate voltage could get you down into the sub 5K ohm range (where you could actually get a transformer made) at a "reduced" power level. The 3-500Z has a very high Mu (160) and this goes against the current trend toward low Mu tubes. That does not mean that it is impossible, just different. I have seen some decent amplifiers built with 811A's, but they took a lot of tinkering to get right. High Mu tubes generally have a higher plate resistance, and are not as linear as their low Mu brethren.
I had a pair of 4PR1000's. I thought about it for about 5 minutes, and then sold them on Ebay when the reality of the cost and size of building an audio amp with them set in. The filament transformer alone is bigger than some amplifiers that I have built, and probably costs more.
Any way that you look at it a high power vacuum tube amplifier is not cheap. The prices go up quickly if you are doing something that is not mainstream. Custom transformers are expensive, and as I found out, not a guarantee of success. If this is your first audio design, it may be wise to build an amplifier that can be built with off the shelf components unless you can afford to experiment a lot.
In a single ended design all of the output tube's current usually flows through the output transformer. This means the transformer must deal with a constant magnetic field caused by this DC current. This causes the core to operate at or near magnetic saturation, which will not be useful for linear coupling. To combat this the core contains an air gap and 5 to 10 times more iron. As transformers get larger it becomes harder to get a wide enough bandwidth for the entire audio range. The exception to this is a "parafeed" circuit which couples the DC into the output tube through a choke, and the audio is extracted through a capacitor into the output transformer (similar to an RF amp). A push pull transformer can be used in this application.
I have been told that the upper limit for a practical SE transformer is 50 to 75 watts. I had a custom transformer made for the 833A at the 75 watt level. It was not cheap. I tested it in a prototype design. It did not work well enough for HiFi, but it will make a killer guitar amp. It can easilly pass 200 watts over the guitar frequency range. The details can be found here:
http://www.tubelab.com/833SE.htm
There is a Yahoo group for people contemplating "extreme" amplifiers:
http://groups.yahoo.com/group/GM70/
You must register and be approved to read the forum. It takes a few days. I asked about BIG SE output transformers and got the same two answers that I have heard here. The Hammond 1642SE (75 watts) and the Electra-Print CU5KB (50 watts). The Hammond weighs 28 pounds, so we are talking about a large, heavy amp. I have not talked to anyone who has actually tried one in a high powered amp though, and some Hammond SE transformers have a less than stellar reputation. Electra-Print transformers are top quality, and they do custom transformers. They are not cheap. I asked about custom 100 watt SE transformers for the 833A and was told that a pair would be $1000, and that was almost 2 years ago. Transformers have gone up in price.
If you are looking to build a 100 to 200 watt amp I would consider a push pull design. The typical amp in this power range uses paralleled audio tubes like the 6550 or KT88. Four to six tubes per channel is common. This requires a low plate to plate load impedance and off the shelf transformers are available up to 280 watts. I have also used TV sweep tubes to make high power. 4 6LW6 (or similar sized) tubes per channel will make an easy 250 watts with the same transformers.
I understand the fascination with big transmitter tubes, so I would use a pair of 813's in push pull. These could do an easy 400 watts at high plate voltages, but again you are looking at a custom transformer due to the high impedances (20K ohms) and power levels involved. 2 or even 4 813's per channel at lowish (1KV) plate voltages and might work with available transformers. There are other possibilities like a pair of 211's, 811A's GM70's or similar tubes.
I am not sure what could be done with the 3-500Z's but a pair of them could provide enough power to set your speakers on fire if a proper transformer could be found. The data sheet shows 1420 watts in class B with an 8.6 K ohm transformer at 3KV. Running in class AB with a lower plate voltage could get you down into the sub 5K ohm range (where you could actually get a transformer made) at a "reduced" power level. The 3-500Z has a very high Mu (160) and this goes against the current trend toward low Mu tubes. That does not mean that it is impossible, just different. I have seen some decent amplifiers built with 811A's, but they took a lot of tinkering to get right. High Mu tubes generally have a higher plate resistance, and are not as linear as their low Mu brethren.
I had a pair of 4PR1000's. I thought about it for about 5 minutes, and then sold them on Ebay when the reality of the cost and size of building an audio amp with them set in. The filament transformer alone is bigger than some amplifiers that I have built, and probably costs more.
Any way that you look at it a high power vacuum tube amplifier is not cheap. The prices go up quickly if you are doing something that is not mainstream. Custom transformers are expensive, and as I found out, not a guarantee of success. If this is your first audio design, it may be wise to build an amplifier that can be built with off the shelf components unless you can afford to experiment a lot.
well I think I've kinda changed my mind on the power tube I'm going to go with. the 813 is looking alot better to me now because of price but it has a much lower drive requirement.
Though how do you know when this will be a problem. I have to say I though I knew alot but i am feeling a little bewildered right now.
Though how do you know when this will be a problem. I have to say I though I knew alot but i am feeling a little bewildered right now.
Tubelab thats my big problem I have all these big transformer on hand. I've got this big 2500v 2kva hyperasil transformer just asking to run some big amp and I don't want to play anymore with big high power rf amp's.
Oh by the way does anyone need a 10 kilwowatt klystron and or a 2.5 megawatt magnetron I'll sell'em cheap
I even have a filment transformer that is isolated up to 40kvdc 10volts at 20 amps but it's larger then most amps i've seen on here.
Damn I hate being obsessed with power it's expensive
Oh by the way does anyone need a 10 kilwowatt klystron and or a 2.5 megawatt magnetron I'll sell'em cheap
I even have a filment transformer that is isolated up to 40kvdc 10volts at 20 amps but it's larger then most amps i've seen on here.
Damn I hate being obsessed with power it's expensive
The other nice thing about the 813 is that it's been used in this application before, so you can find some design hints, guidelines, and don'ts.
A little searching on the forum will get you quite a few book suggestions to bring you up to speed in the audio tube world. My personal recommendations are the Radiotron Designers' Handbook 4th edition (very dated, but very detailed and comprehensive) and Valve Amplifiers 3rd edition (much more modern and superbly useful, but less comprehensive than RDH4). RDH4 is available for free download; Valve Amplifiers is available at Amazon. I'd get both as a start.
A little searching on the forum will get you quite a few book suggestions to bring you up to speed in the audio tube world. My personal recommendations are the Radiotron Designers' Handbook 4th edition (very dated, but very detailed and comprehensive) and Valve Amplifiers 3rd edition (much more modern and superbly useful, but less comprehensive than RDH4). RDH4 is available for free download; Valve Amplifiers is available at Amazon. I'd get both as a start.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.