No element should be floating in a tube. This can create charged islands.
Make a "jumper" connection that grounds pin 9
Make a "jumper" connection that grounds pin 9
Added the jumper to pin 9
PCB layout is attached. Please let me know if all looks good. Anything that can be improved? Or can cause hum?
PCB layout is attached. Please let me know if all looks good. Anything that can be improved? Or can cause hum?
An externally hosted image should be here but it was not working when we last tested it.
Last edited:
I soldered a board and I am testing right now. B+ I am using is about 137v. I am testing with a resistor R21=4.7K and not the CCS (schm in post 20).
Most of the measurements look ok. I get about 40.25v drop across R21, which gives me about 8.5ma of current. Cathode is at 42v, Vgk measured is about -1.725v. plate-cathode is about 93v. So far so good, but then I measure the grid and it sits at about 27v instead of about 40v, but Vgk measures correctly. I dont see any significant drop across R14 or r16 either. Where did I loose these 13 volts.
Most of the measurements look ok. I get about 40.25v drop across R21, which gives me about 8.5ma of current. Cathode is at 42v, Vgk measured is about -1.725v. plate-cathode is about 93v. So far so good, but then I measure the grid and it sits at about 27v instead of about 40v, but Vgk measures correctly. I dont see any significant drop across R14 or r16 either. Where did I loose these 13 volts.
The 1.5nF and 10k Ohm present a 7k Ohm load at 10kHz.
Was the intent to drive 1.5nF of shielded cable?
By the way, the 1 Meg Ohm you have is the maximum resistance rating for the grid return resistor.
Was the intent to drive 1.5nF of shielded cable?
By the way, the 1 Meg Ohm you have is the maximum resistance rating for the grid return resistor.
This is a common occurrence. It is due simply to your meter loading the grid and altering the bias point. You cannot normally measure grid voltage directly for this reason. Just measure the drop across the cathode bias resistor.
Cheers
ian
The time constant of the 4.7uF cap and 1Meg Ohm at the output is 4.7 seconds.
That is very (too) long.
It can go from 30V to 10.9V in the first time constant (4.7 seconds).
What ever is at the output will receive lots of volts until the 4.7uF cap charges to 30V.
What is the DC resistance of the load at the output (like a power amp input for example).
Can it stand 30V going to its input, and then reducing to 10.9V in 4.7 seconds, and 5 times 4.7 seconds (23.5 seconds) going to 1.5VDC.
The frequency response of 4.7uF and 1Meg Ohm is -3dB at 0.03Hz.
If the next stage has 10K Ohm input, the low frequency response is -3dB at 3.4Hz.
That is very (too) long.
It can go from 30V to 10.9V in the first time constant (4.7 seconds).
What ever is at the output will receive lots of volts until the 4.7uF cap charges to 30V.
What is the DC resistance of the load at the output (like a power amp input for example).
Can it stand 30V going to its input, and then reducing to 10.9V in 4.7 seconds, and 5 times 4.7 seconds (23.5 seconds) going to 1.5VDC.
The frequency response of 4.7uF and 1Meg Ohm is -3dB at 0.03Hz.
If the next stage has 10K Ohm input, the low frequency response is -3dB at 3.4Hz.
1.5nf is the worst case capacitive load including cables and connectors, rf filter at the amp input and miller capacitance of the amp.
The grid leak resistor is the 470k. 1Meg is at the input in case of pot skipping, no input connected situations. Probably I should increase the 1meg as its not large as compared to the 470k.
Thanx, thats what I was looking for.
The 1meg at the output is there so that if no load is connected, the output capacitor end is not floating. When you connect a 10k amp, the 1meg has no effect (1meg||10K is almost 10k). Yes, the -3db low frequency response is 3.38Hz. A 10uf is probably a better choice for 10k amp and 4.7uf for 20-25k.
Just be sure to warm up the preamp for a minute or so, before you connect it to the next amp. The cathode has to come to its final DC value, the 4.7uF cap has to charge (with the aid of the 1 Meg resistor). That takes a lot of time.
Otherwise, you will have 30V, 20V, or perhaps 10V applied to the 2nd amp input. Most amps do not like this. And if the 2nd amp is already on, it will put a loud transient to the loudspeakers.
Otherwise, you will have 30V, 20V, or perhaps 10V applied to the 2nd amp input. Most amps do not like this. And if the 2nd amp is already on, it will put a loud transient to the loudspeakers.
I ran some measurements on the preamp with my pc and asus essence soundcard. Pictures are attached below.
I am still not using the shunt supply. I faced and issue where the output voltage was not rising on the supply. I badly wanted to try the preamp, so I cut two traces just below the supply in and out bulk caps and soldered a 1k resistor between the +ve terminals of the two caps at the bottom of the board. Later I noticed that R30 wasnt soldered properly. LTspice confirmed my suspicions and showed the same behavior. Anyway, right now its just a crc supply with 220uf-1k-220uf till I fix the shunt. You can see that the 100hz ripple in the images is about -90db. Will post results again once I fix the shunt regulator.
Baseline with asus essence st sound card -
This one is using a CCS load running at about 9.5ma.
This one is using a 4.7K load resistor instead of a CCS
The CCS seems to make quite a lot of difference, the second harmonic is down from -67db to -77db and the third goes from -97db to -112db.
I also tested with ARTA, seems to gives very slightly worse numbers, rest all looked quite similar so I wont post those.
I am still not using the shunt supply. I faced and issue where the output voltage was not rising on the supply. I badly wanted to try the preamp, so I cut two traces just below the supply in and out bulk caps and soldered a 1k resistor between the +ve terminals of the two caps at the bottom of the board. Later I noticed that R30 wasnt soldered properly. LTspice confirmed my suspicions and showed the same behavior. Anyway, right now its just a crc supply with 220uf-1k-220uf till I fix the shunt. You can see that the 100hz ripple in the images is about -90db. Will post results again once I fix the shunt regulator.
Baseline with asus essence st sound card -
An externally hosted image should be here but it was not working when we last tested it.
This one is using a CCS load running at about 9.5ma.
An externally hosted image should be here but it was not working when we last tested it.
This one is using a 4.7K load resistor instead of a CCS
An externally hosted image should be here but it was not working when we last tested it.
The CCS seems to make quite a lot of difference, the second harmonic is down from -67db to -77db and the third goes from -97db to -112db.
I also tested with ARTA, seems to gives very slightly worse numbers, rest all looked quite similar so I wont post those.
I liked the sound of this ecc88 version quite a lot. I did most of my listening with the CCS version. The sound is extremely clean and quite musical. Initially I thought its quite like any good SS preamp, but after listening some more, the liveliness of music becamse quite evident. I should stop now as I am not much good at describing music.🙂
Now I want to make a version with a 12v tube and this one I want to make an effects box, with loads of harmonic distortion. I have a few tubes for this, tried the 12ax7 in ltspice and plate curves, but didnt find a satisfactory good operating point, this dude likes high voltage and I only have a 110v transformer, gives me about 165v dc and after dropping some 25-30v in the crc filter, I only get about 135v.
Next I tried 12au7 and I think I liked this one. I probably wont bother with a ccs with this one or the shunt siupply. The current consumption is less than half of the ecc88 version, so the ripple will be better than that also.
this is the operating point I chose.
Schematic and sims -
Now I want to make a version with a 12v tube and this one I want to make an effects box, with loads of harmonic distortion. I have a few tubes for this, tried the 12ax7 in ltspice and plate curves, but didnt find a satisfactory good operating point, this dude likes high voltage and I only have a 110v transformer, gives me about 165v dc and after dropping some 25-30v in the crc filter, I only get about 135v.
Next I tried 12au7 and I think I liked this one. I probably wont bother with a ccs with this one or the shunt siupply. The current consumption is less than half of the ecc88 version, so the ripple will be better than that also.
this is the operating point I chose.
An externally hosted image should be here but it was not working when we last tested it.
Schematic and sims -
An externally hosted image should be here but it was not working when we last tested it.
Last edited:
I posted the ECC88 measurements, those are in the last post at the previous page.
I am getting the 2nd harmonic about -77db with the ccs, third is -112db which is ok. Can this be improved. I thought this tube can do THD of 90db easily with a ccs.
I am getting the 2nd harmonic about -77db with the ccs, third is -112db which is ok. Can this be improved. I thought this tube can do THD of 90db easily with a ccs.
- Status
- Not open for further replies.