Building Big Tube amps has been a pet fetish for me since the 1970's when I was working on medium sized monsters (100 watt-300 watts) in my job. The problem with even a small 300 watt amp was the size of the transformers ( cost and availability is a whole other issue ).
I worked on a Phillips 300 watt PA amp in the late 70s which used two QB3/300 TX tubes in the OP stage. It ran on 2.8 KV and the power and output transformers were around 12 inches cubed. It was made on 3 chassis in a large rack cabinet ( that took 4 guys to lift - or two men and a big trolley ) and it ran 100 volt speaker lines into large metal horns throughout a large factory. I recall when I repaired it, the very next day, while on soak test, the output transformer caught on fire and it was site to see. It had spark gaps and large Oil filled capacitors and two large mercury rectifiers. It ran off a 15 amp mains at 240 volts ac that is 3.5KW available at the input, here is Australia.
So as no one will build big transformers for Audio – Mind you I have tried and tried and the legal liabilities are what scares some winding manufacturers off - A simple way to get the Kilowatts out of off the shelf parts is to use the following process. Again cost is an issue, but its a more practical construction process and the lesser of two evils –unless you keen to design and wind your own iron.
Make a number of power output stages, say 4 in all.
Drive all the output tubes from a common driver phase splitter combination - and with each output stage, drive a separate output transformer.
In this case you will have 4 output transformers. You then apply the secondaries in series parallel, or just parallel if you desire. Feed back can easily be applied if required and the tubes are common as grass to buy. For me, I favour 807’s.
So four 100 watt output stages will deliver 400 watts in all combined into a single speaker. If you run them as 4 separate outputs into 4 speakers, you can set it up so if one stage dies the other three still operate and you don’t go out of operation altogether. You might use 8 tubes, or 16, - it’s your choice. Tube matching is ideal, but not essential.
This concept was used in the early days of Tx modulators - paralleled output stages driving one massive transformer or two smaller output transformers.
RCA and Westinghouse used this approach with good results.
If you use Class AB1 which has just drive voltage swing on the input grids, but no grid current flowing - your driver stage will be much easier to manage and design. You may use more tubes to get the big watts, but as they are cheap so what. If properly designed they will last longer as well. Keep the plate dissipation to around 60% and get rid of the heat.
My own concept amp is looking at maybe 32 x 807 tubes running in the 56 watt AB1 mode per pair. That is nearly 900 watts output, or 224 watts per amp, if I have 4 output stages. I can do it with 4 smaller 250 watt transformers or do the overkill with 4 Sowter 400 watt Hifi Output Trannies.
Also too with the power supply - use stacked low voltage supplies where possible. Easier to source smaller trannies in bulk buys.
If you apply this approach say with 807 tubes you can get the above power figures from 4 power stages with 32 tubes on a B+ of 600 volts, and a half supply of 300 for the screens; and almost no grid drive power.
To do it with two big tubes you might require 2.5 to 3 kv for the plate volts and a very, very expensive non-hifi set of transformers to boot. Most BIG tubes also need driver transformers as well, since many are set up for class B. Smaller tubes not so much. Big tubes also need a decent driver stage which maybe another power amp of 25 watts or so.
Everything being equal – the above approach is a compromise, so you have to decide what is the best for your needs.
But for those that insist of BIG Iron and BIG Tubes -and some serious material on designing high power Audio Transformers, I can recommend the GE/ Naval publication from the 1960s called – High Power, High Voltage, Audio Frequency Transformer Design Manual. Almost 700 pages of material on designing audio transformers from 10kw to 350 Kw in size. A rare book indeed.
If you really want to build an amp that large then you are pretty much on your own. Design using parts that are available otherwise you will spend a great deal having custom transformers wound. I'm using a cnc machine power transformer reverse wired for the plate supply for 20 kt-88s. Output transformers are hammond 280 watt units. 4 of them, 2 in parallel for each channel. Also have to have something to drive all those output tubes and I'm using mosfets, one for each power tube. Those have to have their own supplies along with the filament voltage, regulated preamp filament voltage, regulated preamp high voltage, regulated output tube grid 2 voltage, 12V for cooling fans, 5V for digital control circuitry, bias supplies, etc. It's a lot and I designed custom circuit boards as opposed to point to point for the preamp, power amps, preamp controller, and some of the power supplies. I'm coming down the home stretch now, just dialing in the preamp then need to buy power tubes, test the power amps, and build cabinets. I've been working on this off and on for 3 years and have spent over 3 grand on this project and I expect it to weigh about 250 lbs. It'll be in two separate cabinets- power supply and amp. Here's a sneak peek...
The way professionals have been handling this problem for years is simple. If you need (say) 400W then you get four 100W amps and mount them all in a big rack mount frame. Then you typically have four big 12 inch drivers in the speaker cabinet so they run an 8-pole speaker cable that connects each of the four drivers to it's own amplifier. With solid state class D amps the numbers just get bigger but you see the same set ups with 8-pole Speakon terminated cables.
My chassis were custom made of .125 aluminum, tig welded, powder coated and screen printed with the custom graphics using a two part white epoxy. I designed the graphics which took forever to get everything all lined up with the holes in the chassis. They were not cheap!. The panel that holds the output transformers is .25 aluminum u channel.
You will like having 150 output tubes per side. The turn on thump alone would shoot your speaker cone out like a bullet, into the wall.
How many tubes would it take to drive an 8 ohm load to 500W with the OTL design? OK maybe go with a 16 ohm load
I think the only way to go other than making a custom and very expensive transformer is to use many standard output transformers in parallel. Then the question comes up as to how to combine their outputs. The usual solution is to add the output in the air. Back in the days where all there were was tubes that was how "everyone" built kilowatt sound systems. The prime example was the very large Hamon organs installed into large churches. This instruments use 100% tubes and would literally shake the walls of a large building. Typically the amps were instaled inside cabinets. Three amps and several speakers in each cab. Then they would add cabinets until they had the desired sound volume level. The amps were not large but all those drivers would creat some impressive sound.
Yes I'm paralleling both and it will be both instrument and hifi. It has line inputs as well as instrument inputs but I designed it primarily for bass guitar amplification which needs a good low frequency response. It has a stereo preamp. One channel with a baxandall style tone stack, the other with a fender bassman style tone stack. The line ins bypass the tone stacks and either preamp channel can be assigned to either power amp and line out via the front panel or foot switch.
To produce that kind of power would not take that many 6C33s if you ran OTL. Also, the 'fixed output impedance' thing is not a concern.
We built a 500-watt OTL using 6AS7s: Atma-Sphere music systems, inc. That took only 42 tubes per channel, you could make 900 watts with 90 6AS7Gs. The 6AS7G is still being made.
Think about this for just one second. Each 6C33 draws 6.5 amps of heater current. 150 of them is about 1000 AMPS. That's 6 KILOWATTS just to light the tubes. Game over!
I have experimented with this method. I have a large quantity of guitar amp OPT's that are good for anywhere from 25 to 100 watts depending on the driving impedance, load and required low frequency response. I have a succesful EL84 based amp that makes 60 WPC using a common driver for two push pull pairs of EL84's per channel running from 430 volts B+.
Simply paralleling two OPT's results in the effective primary inductance dropping in half. This is OK if the number of tubes is also doubled.
As previously stated I ran a seperate pair of output tubes for each transformer with no connection between the primaries of each OPT. All tubes ran from a common driver with individual bias adjustments. The best results occur with the secondaries wired in parallel, but this often results in a low output impedance if several sets of tubes/OPT's are paralleled.
Wiring the secondaries in series can result in fireworks if one set of tubes is weak and not sharing the load duties. The output current from the other OPT's is forced through the secondary of the weak set, and which can result in a large voltage drop across the secondary, and a huge voltage induced in the primary. Cranking a guitar through a setup like this at levels well into clipping can and WILL invite a visit from the fire gods.
Further experiments are planned in the future, but my tube time has been rather limited now, so I can't say when. If I get anything working soon, and this thread is still active, I'll try to remember to post it here. Otherwise watch this one:
http://www.diyaudio.com/forums/tubes-valves/203110-mono-push-pull-driver-pcb-13.html
Previous big powered stuff (525 watts paralleled mono) was here:
http://www.diyaudio.com/forums/tube...new-p-p-power-amp-design.html?highlight=35lr6
Let me calculate please... I have 10 fingers on hands.... No more, sorry,,, Let's calculate using 10 fingers... Each tube wants 5K. The typical boutique fancy horn has 8 Ohm in the middle of it's waist. How many of tubes would I need? It is doable. But who wants to pay for such an experiment?
The OP originally asked about making a 500 WPC tube amp. Reasonable cost and performance goals were assumed. Later, reality convinced him to look at 200 WPC, then 120 WPC. I don't know what happened to him, but he hasn't been seen on this forum in about 8 months.
It is common knowledge that the transformers in a tube amp represent a big portion of the total cost. In a BIG tube amp that portion could be like 90%. I don't claim to be an OTL expert, but I think a 500 WPC OTL would be big, and produce considerable heat. So for reasonable cost, performance, and thermal load (I live in South Florida it's hot enough), I will assume the use of one or more OPT's per channel.
It has been suggested that multiple small OPT's could be combined to produce a big amp. I have experimented with this, but on a small power scale. Let's scale things up and see if we make big power or summons the fire gods! How big?????
The OP wanted 500 WPC, then 200, then 120. Well I did a 125 WPC amp, and several builders succesfully cloned it. Using Ebay sourced power transformers, and Edcor OPT's the total iron cost is just over $200. Using Edcor OPT's and Antek toroids for power makes that $300, add $150 for everything else.
So lets think BIGGER. The new target was 200 WPC, and the "goal" is 500 WPC! OK, how do we get there.....in other words, how many tubes and transformers do we need to wire together. Lets not calculate please.....lets just wire things up and turn up the juice! OK, lets guess how many to wire up, and what to connect them to.
Recently I have been connecting things up to one of Pete's driver boards. I built that board so that it could drive just about anything, and have recently tested it with a few tubes. The real surprise came when I connected up some ultra cheap 13GB5's and one of the guitar amp OPT's that I used with the EL84's. I could squeeze 150 watts from a pair on 600 volts. See post #66 here:
http://www.diyaudio.com/forums/tubes-valves/203110-mono-push-pull-driver-pcb.html
So, I propose a simple experiment, wire up 4 pairs to 4 OPT's, get the fire extinguisher ready, and turn the power supply UP!
First I will make measurements on one set at power levels up to 125 watts, then add 3 more sets, and test again. I can connect a real Plitron for comparison, but the comparison wouldn't be fair....the Plitron cost $400 and the guitar amp OPT's were $20 each!
Sound like a plan????????
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.