Hello everyone!
Has anyone tried a pentode in UNSET topology and loaded with a CCS as voltage amplification stage? I'm getting very low distortion for 160Vpp output voltage, 220K load, gain 36. The gain is controlled by R2/R4 ratio. I want to build it, but first checking whether anyone has played with this. Tried multiple pentodes, even a 6BQ5 gets around 0.08% full swing at 15ma.
Regards,
Jose
Has anyone tried a pentode in UNSET topology and loaded with a CCS as voltage amplification stage? I'm getting very low distortion for 160Vpp output voltage, 220K load, gain 36. The gain is controlled by R2/R4 ratio. I want to build it, but first checking whether anyone has played with this. Tried multiple pentodes, even a 6BQ5 gets around 0.08% full swing at 15ma.
Jose
I have this circuit breadboarded and working quite well with about 20 different tubes. The current record is about 50 volts P-P of output at 0.07% THD with a $1 tube! The residual THD of my oscillator is 0.04%. I know I posted it somewhere, but cant find it now. I did find some of the earlier testing from the day it first came to life. The gain can be controlled by changing R16. See posts # 110, 113, 114, and 119 here:
Attachments
Thank you @Tubelab_com! It all look goods, but then why this arrangement is so rarely used to drive big tubes like 300B? It seems capable of easily delivering 140Vpp, low distortion etc.
This circuit was created by combining the UNSET topology with a bootstrapped load pentode circuit that I had used in a guitar amp design about 10 years ago. I called that circuit the SATURATOR since any common electric guitar could easily drive it into full saturation. I got the brilliant idea to combine the two thinking that it would be a neat feature for adjustability in a guitar amp. R16 is the gain or "saturation knob" and R9 affects the slope of the "triode" curves created by the "Schade" feedback. It does work well in that application.
Until I built and fully tested the circuit, I had no idea that it works so well as a ultra low distortion driver for just about anything. I have already got a "ratsnest build" version of an UNSET amp that makes an easy 25 WPC with paralleled 26HU5's for outputs. The 300B experiments are ongoing, but looking good. Before going too far down this road I need to settle on a good choice for the driver tube, or tubes.
About 20 years ago I picked the 5842 tube for the TSE and TSE-II amps because it WAS cheap, and it outperformed everything else I tried. Unfortunately, TSE builders have made the 5842 rather scarce and expensive today in some parts of the world.
I have seen the bootstrapped load by mosfet circuit called a gyrator in several places including many posts on diyAudio, and it does have some of the properties of an inductor. The traditional opamp circuit is the common choice for a "gyrator" but opamps that eat 300 volts are not common.
The low frequency end is directly related to C2 / R16. R16 also controls the gain, so there is some interaction, but getting below 20 Hz is not an issue. The HF is affected by the mosfet capacitances. Large mosfets used in this LT Spice sim are not your friend. More reasonable sized parts get you well above the audio range except at very high gain settings.
The OPT was still the limiting factor in my 25 watts SE test amp.
Looks great! Someone pointed me to the beauty of the 6ej7 (maybe on hifihaven, I forget) and I am using it with a 10M45 ccs as input stage of my 300b amp. It sounds great. I have been itching to complicate that input stage a little bit.
8.1V. Plenty of choice, I'd go for something with a low Crss capacitance, 100V max Vds. It is a source follower, should not be very critical. V4 voltage will depend on the mosfet threshold, therefore it needs to be adjustable.
I added a more realistic CCS to see how it affect the distortion and bandwidth. Distortion stays low, and bandwidth shows a bump at after 100Khz, but nothing drastic. This is all, of course, simulation, but so far it makes sense.
I have not put any protection to the mosfet gate etc, this would need to be added for a real circuit, of course.
I added a more realistic CCS to see how it affect the distortion and bandwidth. Distortion stays low, and bandwidth shows a bump at after 100Khz, but nothing drastic. This is all, of course, simulation, but so far it makes sense.
I have not put any protection to the mosfet gate etc, this would need to be added for a real circuit, of course.
You can even replace the pentode with another 10M90S, but then this would belong to a completely different forum 🙂
If someone somewhere would actually replace the 6EJ7 with a mosfet it might look something like this. It might even have it's own thread here:
My involvement starts in post #63 with a familiar schematic in post #87. I
A certain Big Dumm Blond tinkerer might have obtained a board or two, built one and tested it. The 0.233% THD at 30.0 Vrms output might have prompted him to hook it up to an UNSET output board where the combo made 25 watts at 2.27% THD and 0.311% THD at 1 watt in SE mode with a 26HU5 tube.
Those results might have tempted him to defect totally to the dark side for a week or two. FETSET might have been created THE SCG board is seen here driving a perfboard creation that has a IXYS IXTH6N100D2 in place of the 26HU5 tube. The schematic for the whole FETSET thing is included.
The Big Dumm Blonde One tried to make it bigger with paralleled output mosfets. This was the bridge too far which would crank out 40+ watts for hours them just randomly blow up. The remnants of those experiments might be sitting in a box begging for some attention.
It seems that a certain breadboard that had three small mosfets and two big ones per side that might have matched the included schematic is missing from that box. Could it be mounted on a heat sink and wired to a pair of speakers somewhere? I'll never tell.
My involvement starts in post #63 with a familiar schematic in post #87. I
A certain Big Dumm Blond tinkerer might have obtained a board or two, built one and tested it. The 0.233% THD at 30.0 Vrms output might have prompted him to hook it up to an UNSET output board where the combo made 25 watts at 2.27% THD and 0.311% THD at 1 watt in SE mode with a 26HU5 tube.
Those results might have tempted him to defect totally to the dark side for a week or two. FETSET might have been created THE SCG board is seen here driving a perfboard creation that has a IXYS IXTH6N100D2 in place of the 26HU5 tube. The schematic for the whole FETSET thing is included.
The Big Dumm Blonde One tried to make it bigger with paralleled output mosfets. This was the bridge too far which would crank out 40+ watts for hours them just randomly blow up. The remnants of those experiments might be sitting in a box begging for some attention.
It seems that a certain breadboard that had three small mosfets and two big ones per side that might have matched the included schematic is missing from that box. Could it be mounted on a heat sink and wired to a pair of speakers somewhere? I'll never tell.
Attachments
There are probably lots of things that I haven't done either because I didn't think of it yet or didn't have the parts / material to try it.
I have been tinkering with electronics ever since I started dragging vacuum tube powered table radios home as a young kid. Dead radios turned into dead TV's and a lucky find of three identical Sylvania TV's gave me enough parts to make one working TV. There weren't too many kids that had their own TV in the early 60's. I had the only color TV in the neighborhood in 1971 when one of the Apollo missions broadcast live color TV from the moon, which was replayed daily for a week or so. It was a 1957 vintage Emerson that I had rebuilt with an all glass CRT from a Philco that was replaced due to the creping crud that grew in from the sides on some of the old round tubes in the 60's.
With no formal electronics education I showed up at a Motorola plant in response to a "technicians wanted" ad in the Miami Herald newspaper. Over 500 people showed up for two jobs. Somehow my "I can fix things, hire me" story and the second highest score ever made on their electronics test did the trick. 41 years later I retired as a "Principal Staff Research Engineer." During that time Motorola paid for me to get two engineering degrees.
They also gave us free silicon for "home projects" if you had a good story on the sample request form. Finding out that the in house sales engineer was a ham radio guy and helping him with his antenna might have greased the skids a bit. Fixing an electron microscope at exactly the right moment in my career nailed the free stuff for life thing. A few of us who did DIY audio stuff made clones of every SWTPC Tiger except the Tigersaurus. We cloned the SWTPC 6800 computer too. I got samples of chips that didn't even exist yet! Note the complete absence of part numbers on the CPU chip in the board on the left. I don't think that one ever made it to production as flash memory came along about the time this would have appeared, and the old 8 bit HC11 architecture was retired.
Attachments
Very impressive resume! Unfortunately I can only relate to the tube scavenging from old TVs bit in my teenage years. Growing up in Cuba everything tech related was a challenge, I had to borrow soldering gun, multimeter etc from a neighbour who repaired radios and TVs. Then studies took me in a different direction, did not study or worked in electronics, but the hardware bug was always present.