My college age son wanted to try his hand at building his own amp. After a lot of searching and thought we settled on this amp. It's been a true father / son experience. He's got the board filled, we've built a chassis and mounted everything. Not surprisingly he wants to "plug it in" to see how it sounds. I've convinced him to slow down, and we've already found some mistakes.
Anyway, it was fun to watch the excitement and remember when my first urge was to plug it in and see what happens.....
Anyway, it was fun to watch the excitement and remember when my first urge was to plug it in and see what happens.....
Very cool.
When I was a little kid (maybe 8 or 9) I always used to get my mom to plug "it" (whatever my creation was) in for me. Just in case there was, um, a wiring issue
No safety worries back then. I remember the line-voltage hot dog cooker (two 16-penny nails driven through a board, each connected to one wire of a line cord - worked great!), and making my own carbon arc lamp (two carbon rods from D-cells, each connected to one wire of the cord, push them together until it arced). Needless to say we went through a lot of fuses in the fuse box.
Pete
Despite my reputation, it was my father, not me, who put the penny behind the AC fuse causing the destruction of the 230 volt outlet.
My parents were afraid.....very afraid, of some of my creations.
I used the big fat ones from a glow plug (model aircraft engine) battery. They had a screw terminal already attached. Wire them in series with a portable space heater for a ballast.
Last edited:
I *know* this will turn out to be a question with an obvious answer, but I would rather be safe than sorry.
I'm prepping to build the DCPP amp and I can't quite grok the orientation of the tube sockets for the inverted build (sockets being on the non-silk-screened side). It seems to me that if I were to solder these sockets in inverted they would be backwards (pin 7 in hole 1; pin 1 in hole 7) for the 6CB6. Yet I've seen pictures of both non-inverted and inverted builds with no extra leads to reverse the ordering.
What's going on?
This is my first tube build, so maybe there's a pin-numbering convention that I'm missing?
Thanks.
I'm prepping to build the DCPP amp and I can't quite grok the orientation of the tube sockets for the inverted build (sockets being on the non-silk-screened side). It seems to me that if I were to solder these sockets in inverted they would be backwards (pin 7 in hole 1; pin 1 in hole 7) for the 6CB6. Yet I've seen pictures of both non-inverted and inverted builds with no extra leads to reverse the ordering.
What's going on?
This is my first tube build, so maybe there's a pin-numbering convention that I'm missing?
Thanks.
ahankinson,
The "normal" build will have the tube sockets on the opposite side of the circuit board from the rest of the components. This is the way Mr. Millett designed it so, just install the sockets that way, and it will work as intended.
I think I know which photo you are talking about ( with sockets on the same side of board as the components). That is the one that George at TUBE LAB used for his testing. I don't have a tube pin out chart in front of me at the moment but, I'm fairly certain that if you install the sockets on the silk screened side of the board, you WILL have to do some modifications of the board by cutting some traces and running jumper wires.
Just install the sockets on the opposite side as the components/ silk screening, and it will work as designed.
Hope this answers your question.
Dan
The "normal" build will have the tube sockets on the opposite side of the circuit board from the rest of the components. This is the way Mr. Millett designed it so, just install the sockets that way, and it will work as intended.
I think I know which photo you are talking about ( with sockets on the same side of board as the components). That is the one that George at TUBE LAB used for his testing. I don't have a tube pin out chart in front of me at the moment but, I'm fairly certain that if you install the sockets on the silk screened side of the board, you WILL have to do some modifications of the board by cutting some traces and running jumper wires.
Just install the sockets on the opposite side as the components/ silk screening, and it will work as designed.
Hope this answers your question.
Dan
The sockets MUST go on the non silkscreened side, otherwise they will be wired wrong. The board is designed this way.
If this is your first build. I suggest that you build it the way Pete intended. It works great, and the sockets are on the opposite side of the board from the other components.
If you are referring to all of the early pictures in posts with my name on them, The sockets ARE on the side without silk screen. I also put all the components on the side without silkscreen. I did this since I got the board for experimenting and never intended to put it in a case. It is just easier to work on a board and change parts when everything is on the same side.
I do have two other DCPP boards that were built in the normal fashion and one does spend some time in a chassis....but I still get the urge to take it apart from time to time. In fact I took the board out and loaned it to my boss so he could look at it while building the board he bought. None of the Tubelab amps would drive his Manepans.
The toaster or space heater is the key to saving fuses, a cheap ballast with a positive tempco. It also keeps the carbon rods from exploding.
Today, I have an old HP power supply that is seen in many of my pictures of Petes big red board. It makes 650 volts at 1.7 amps turned all the way up. The current limiter is very slow. Slow enough that the supply will vaporize a 47uF 450 volt cap leaving only a black stain and two leads sticking out of the red board.
To use this power supply for "destructive testing" I use two 8 ohm 500 watt resistors in series with the device under test. Too much "testing" trips the 15 amp bench breaker.
Back on thread....I'm currently wiring up a big red board for my boss, the 125WPC version to rock his Maggies. Maybe he will post pictures after he makes the cabinet.
Plate voltage too high
I'm sure this is shown somewhere in this thread but I can't seem to find it. I've finished my build and it sounds fine but I am not getting much volume. I am trying to find the normal readings for the 6CB6 tubes because my plate voltages are too high. Here is what I am getting.
B+ 330V
Pin readings to ground
Pin 1 0.2mV
Pin 2 1.797 V
Pin 5 215.5V
Pin 6 140.0 V
Pin 7 1.795V
My C- is -61V
Heaters are 5.95V AC accross pins 3 & 4
Thanks for your help!
I'm sure this is shown somewhere in this thread but I can't seem to find it. I've finished my build and it sounds fine but I am not getting much volume. I am trying to find the normal readings for the 6CB6 tubes because my plate voltages are too high. Here is what I am getting.
B+ 330V
Pin readings to ground
Pin 1 0.2mV
Pin 2 1.797 V
Pin 5 215.5V
Pin 6 140.0 V
Pin 7 1.795V
My C- is -61V
Heaters are 5.95V AC accross pins 3 & 4
Thanks for your help!
What are you using to drive your amp? Neither my phono stage or DVD player with drive the amp to clipping, I need to use a preamp with gain.
The one's in my amp all measure ~26.3k, so that's not going to be an issue, as my amp works fine.
jeff
I'm using a CD player through a passive preamp. The same set up drives my ST70 (VTA version) to clipping pretty easily. Just to check I breadboarded a little JFET preamp and that helped a lot, and with that in the mix I can make it clip.
I am still wondering if 200V on the plates is too much. I ordered some more 6CB6's to see if it's just the tubes. The current set are all NOS GE, and I only have 5. I was surprised that switching out one of them dropped the plates from 215 to 200 so I thought it was worth trying.
I am still wondering if 200V on the plates is too much. I ordered some more 6CB6's to see if it's just the tubes. The current set are all NOS GE, and I only have 5. I was surprised that switching out one of them dropped the plates from 215 to 200 so I thought it was worth trying.
You mention the 200 volt plate voltage, and a somewhat low gain. I don't believe these are related.
First off, I went through about 20 well used 6CB6's to find 4 that were somewhat similar in plate voltage. They varied from 150 to 200+ volts. All the tubes will be different, but a small collection of new tubes may not yield any better results. The amp will work fine at 200 volts, and the gain will not vary much with the voltage, but the distortion will. The best distortion numbers seem to be at about 150 volts.
I would set the driver pot in the middle and find two sets of tubes (one set for each channel) that give the closest plate voltage. Then change the value of the 200 ohm resistor in the source lead of the 10M45 to get the plate voltage where you want it. (R39 and R37). The 10M45's seem to vary a lot from batch to batch. I have installed a 500 ohm trimpot in my boards.
If you have distortion measuring equipment, adjust the 200 ohm resistor for lowest distortion, or if you have FFT capability, lowest higher order distortion. Then tweak the "D BAL" pots for distortion (R52 and R54).
That's what you should be reading. The plate resistors are in parallel with the feedback resistors (through the OPT) You should see 220K || 30K = 26.4K
This amp as Pete designed it just barely has enough gain. It just clips with my CD player, but won't clip with my DVD player. When you build the 125 WPC version, nothing will drive it to clipping.
There are a few things that can be done to boost the gain, but all require some experimentation.
I tried several higher transconductance driver tubes to boost the gain. The best one I tried is the 6GU5. It is not a pentode, it is a hexode, but it will plug right into the red board with no circuit changes. It might even bring the plate voltage down a bit, but the 10M45 resistors might require tweaking. The side effect????? Lower overall distortion.
The plate load (R50, R55 and R51, R53) and the feedback resistors (R29, R47 and R30, R31) also play a large role in the overall loop gain of the entire amplifier. The ratio of these resistors controlls the amount of local feedback, while the parallel resistance (the 26.4K mentioned earlier) in parallel with the 100K grid resistor for the output tube (100K || 26.4K = 20.8K), is the plate load resistance the driver tube sees. A pentode will have higher gain with a higher plate load, assuming all other parameters don't change.
You did not say if you are using the GNFB connection this controls the overall gain as well.
If the GNFB is enabled, try the amp without it. I never liked the sound with GNFB, but some speakers need it, and some users prefer it. You can tweak the amount of GNFB with the value of R71 and R72. Higher resistor value is less feedback.
You can lower the local feedback, thus increasing the gain, by increasing the value of the feedback resistors (R29, R47 and R30, R31). All 4 resistors should be the same value. Increasing these resistors will lower the driver plate voltage. Increasing the plate load resistors (R50, R55 and R51, R53) will increas the driver gain, but also increase the local feedback which lowers the output tube gain. The driver plate voltage will also go down. As you can see everything here interacts so some experimentation is needed.
I made up some resistors with aligator clips and experimented considerably. I found that 33K plate resistors, and 240 K feedback resistors worked the best with 6GU5's in an amp with 350 volts of B+ and 36K plate resistors and 330K feedback resistors worked best in 600 volt amps.
First off, I went through about 20 well used 6CB6's to find 4 that were somewhat similar in plate voltage. They varied from 150 to 200+ volts. All the tubes will be different, but a small collection of new tubes may not yield any better results. The amp will work fine at 200 volts, and the gain will not vary much with the voltage, but the distortion will. The best distortion numbers seem to be at about 150 volts.
I would set the driver pot in the middle and find two sets of tubes (one set for each channel) that give the closest plate voltage. Then change the value of the 200 ohm resistor in the source lead of the 10M45 to get the plate voltage where you want it. (R39 and R37). The 10M45's seem to vary a lot from batch to batch. I have installed a 500 ohm trimpot in my boards.
If you have distortion measuring equipment, adjust the 200 ohm resistor for lowest distortion, or if you have FFT capability, lowest higher order distortion. Then tweak the "D BAL" pots for distortion (R52 and R54).
That's what you should be reading. The plate resistors are in parallel with the feedback resistors (through the OPT) You should see 220K || 30K = 26.4K
This amp as Pete designed it just barely has enough gain. It just clips with my CD player, but won't clip with my DVD player. When you build the 125 WPC version, nothing will drive it to clipping.
There are a few things that can be done to boost the gain, but all require some experimentation.
I tried several higher transconductance driver tubes to boost the gain. The best one I tried is the 6GU5. It is not a pentode, it is a hexode, but it will plug right into the red board with no circuit changes. It might even bring the plate voltage down a bit, but the 10M45 resistors might require tweaking. The side effect????? Lower overall distortion.
The plate load (R50, R55 and R51, R53) and the feedback resistors (R29, R47 and R30, R31) also play a large role in the overall loop gain of the entire amplifier. The ratio of these resistors controlls the amount of local feedback, while the parallel resistance (the 26.4K mentioned earlier) in parallel with the 100K grid resistor for the output tube (100K || 26.4K = 20.8K), is the plate load resistance the driver tube sees. A pentode will have higher gain with a higher plate load, assuming all other parameters don't change.
You did not say if you are using the GNFB connection this controls the overall gain as well.
If the GNFB is enabled, try the amp without it. I never liked the sound with GNFB, but some speakers need it, and some users prefer it. You can tweak the amount of GNFB with the value of R71 and R72. Higher resistor value is less feedback.
You can lower the local feedback, thus increasing the gain, by increasing the value of the feedback resistors (R29, R47 and R30, R31). All 4 resistors should be the same value. Increasing these resistors will lower the driver plate voltage. Increasing the plate load resistors (R50, R55 and R51, R53) will increas the driver gain, but also increase the local feedback which lowers the output tube gain. The driver plate voltage will also go down. As you can see everything here interacts so some experimentation is needed.
I made up some resistors with aligator clips and experimented considerably. I found that 33K plate resistors, and 240 K feedback resistors worked the best with 6GU5's in an amp with 350 volts of B+ and 36K plate resistors and 330K feedback resistors worked best in 600 volt amps.
Thank you for the excellent suggestions. I do have the NFB hooked up, and I'll try disconnecting that. I just bought a batch of driver tubes, so I will see if those help too.
I have found that my bias is drifting a LOT, and once it's set I get pretty good output. My DMM, unfortunately, automatically switches right at 400 mV. I picked up a new one today so i'll see if that helps.
I have found that my bias is drifting a LOT, and once it's set I get pretty good output. My DMM, unfortunately, automatically switches right at 400 mV. I picked up a new one today so i'll see if that helps.
Is the GNFB supposed to connected to the common side (white wire) of the OPT? That's what Pete's drawing shows, but the white wire isn't positive, it's common. Thanks for all the good info from me too.
Ok this is a bit weird. I was getting plenty of power after hooking up my JFET preamp. I put my cover on the chassis, hooked it all up and now my power is gone again. I opened it up and checked the bias and balance, and everything is right on.
By the way, GNFB is hooked up, and my driver tubes are pretty close (the balance pot is almost at 12:00 o'clock on both sides).
I haven't tried lifting the feedback. I'll do that next.
By the way, GNFB is hooked up, and my driver tubes are pretty close (the balance pot is almost at 12:00 o'clock on both sides).
I haven't tried lifting the feedback. I'll do that next.
Lifting the GNFB really helped. I also swapped several sets of driver tubes but the biggest difference was removing the feedback. By the way, I found an earlier post suggesting that R71 & R72 be connected to ground if the GNFB is lifted. I did not do that, I didn't understand why you would. Any help??