I haven't been able to get back to this project for several weeks, so you are right, kind of quiet. I will reiterate that the little I've been able to listen to the EO MM, it's totally worth the trip.
One important thought. I have been getting noise problems that are, in part, related to low Idss on the JFETs. I managed to buy Toshiba's that were BL or Idss in the 4 to 6 mA range. If you are buying JFETs, I would recommend getting those in the range of SY's in the article to likely avoid noise issues.
When I get back to it, I will be able to reduce the CCS into the 2 to 3 mA range and, hopefully, get out of my noise issue.
When I get it figured out, I will post other lessons learned in my build.
Good luck on yours. As I said earlier, totally worth it.
Jac
Edit - I just looked up Jim's Audio. Looks like they are the low Idss kind. At that price, the Linear's from the DIY store may be a better deal.
Last edited:
It was shockingly nice here today as well. I was trapped inside at AXPONA hearing some expensive gear that ranged from very good down to "is that on purpose?" awful.
It was a mixed bag. The best sound was a room where they really took a lot of care doing wall treatments- diffusers everywhere, with some spots of absorption. It helped that the speakers were quite good, of course, but there were lots of rooms with what might have been good speakers, but the overall sound was lousy. I did have one earnest young fellow in a cable room "explain" electromagnetic theory to me. I listened attentively and tried hard not to burst out laughing.
Closer to the topic, I met Ralph Karsten (handle here is atmasphere) who is also a strong believer in the efficacy of balanced phono; comparing notes, we found several points of convergence, even though our designs are quite different.
Closer to the topic, I met Ralph Karsten (handle here is atmasphere) who is also a strong believer in the efficacy of balanced phono; comparing notes, we found several points of convergence, even though our designs are quite different.
Ralph seems to have been a bit quiet here recently. I hope its because he's being super successful with the business.
I stopped talking to him after he tried to convince me that feedback does not lower output impedance...
Jan
Jan
Briefly. She was set upon from all sides by typical audio show attendees (I was there with my doctor, whom you met, and his first observation was, "Well, this is quite a sausage fest."), so I didn't want to get in the way of her selling CDs and records.
She will have to take a very long shower to wash off the geek reek.
She will have to take a very long shower to wash off the geek reek.
Quiet At Last!!
@ sonidos,
I figured out my noise problem and it wasn't due to the low Idss of the BL Toshibas. Of course, if you do use the BLs, you will want to decrease the CCS current to around 3 mA to give you a couple of mA between Idss and the CCS current. I also recommend that you use sockets for the CCS control resistors and measure/tune them on the board using the actual power supply. In my case, my JFets are ranging from 5.9 to 6.3 mA and they were the highest pairs I could find. I adjusted my CCS to between 2.7 and 2.9 to have about 3 mA difference.
SY,
Thanks for helping me think about things to look for. I couldn't have diagnosed it without your input on what to look for. I learned a lot about the circuit as well. This experience has reinforced the idea of building and testing on piece of plywood where it's easy to change things.
In the end, I had a bad batch of DN2540 and some of them ended up in the CCS in the original build. What I did was breadboard the CCS to match mosfets. If I had been smart and measured Idss and Vgs(th) instead, I would have caught this a lot earlier. Of the 24 DN2540 I purchased, 8 of them, including the ones that caused the noise, just wouldn't shut off. The drain current would decrease into the 3 mA range at about -2 Vgs, but would then increase as Vgs went further negative. The good ones shut off normally with Vgs(th) in the -1.8 to -2.1 range.
Now, I can stop feeling stupid, listen to a few records, and solder up the things currently on sockets.
Jac
@ sonidos,
I figured out my noise problem and it wasn't due to the low Idss of the BL Toshibas. Of course, if you do use the BLs, you will want to decrease the CCS current to around 3 mA to give you a couple of mA between Idss and the CCS current. I also recommend that you use sockets for the CCS control resistors and measure/tune them on the board using the actual power supply. In my case, my JFets are ranging from 5.9 to 6.3 mA and they were the highest pairs I could find. I adjusted my CCS to between 2.7 and 2.9 to have about 3 mA difference.
SY,
Thanks for helping me think about things to look for. I couldn't have diagnosed it without your input on what to look for. I learned a lot about the circuit as well. This experience has reinforced the idea of building and testing on piece of plywood where it's easy to change things.
In the end, I had a bad batch of DN2540 and some of them ended up in the CCS in the original build. What I did was breadboard the CCS to match mosfets. If I had been smart and measured Idss and Vgs(th) instead, I would have caught this a lot earlier. Of the 24 DN2540 I purchased, 8 of them, including the ones that caused the noise, just wouldn't shut off. The drain current would decrease into the 3 mA range at about -2 Vgs, but would then increase as Vgs went further negative. The good ones shut off normally with Vgs(th) in the -1.8 to -2.1 range.
Now, I can stop feeling stupid, listen to a few records, and solder up the things currently on sockets.
Jac
Last edited:
lehmanhill,
Thanks for posting the challenges you had and how you solved them. Hoping it keeps me out of trouble by following your tips.
Thanks for posting the challenges you had and how you solved them. Hoping it keeps me out of trouble by following your tips.
Wow, that's interesting- I haven't experienced bad DN2540s, other than the ones killed through my stupidity. So now I have one more troubleshooting tool to consider, for which I thank you.
Sy, May I ask why you stayed with tubes when you made the choice to move to Jfets as the input device. Just curious more than anything.
Esthetics. I like tubes, so as part of the design brief, I decided to have all voltage amplification done with them. Theoretically, this also helps with overload margins, since tubes can swing lots of voltage, but I'm not sure this ends up being a real advantage in practice.
Jan and I are currently working on an all-FET version of this (I think innovative) topology; if the performance is just as good as with the tubes, this should be a nice practical way to bring down the size, cost, and energy consumption.
Jan and I are currently working on an all-FET version of this (I think innovative) topology; if the performance is just as good as with the tubes, this should be a nice practical way to bring down the size, cost, and energy consumption.
Usually, I am very lucky with new parts that I buy. I seemed to have attracted a bit of bad part karma on this build. I also had a couple of bad LM329, bad right out of box.
In the future, I think I will buy a few more spares. You can add a few spares that you don't end up using and still save money on shipping fees when you have to go back and order more of something. I will consider the unused spares as offerings to the DIY gods.
Jac
I've never had a bad DN2540 out of the tube, nor have I had a bad LM329. I usually buy 2 or 3 tubes of the former, and 100 of the latter, maybe once a year.
That's the plan, yes. Regulators still on board, I/O set up for terminal blocks. I just need to finish these stupid drawings so that Jan can get the boards...
Wow! You build a lot faster than me. If I did nothing else in my waking life, I don't think I could go through that many in a year.
As for never having a bad DNN2540, I believe you. I certainly didn't expect it 4 of the bad ones were fresh out of the tube and had been treated nicely regarding static. They were MicroChip brand (SuperTex) straight from Farnell (Newark here).
Jac