Technical Assistance/Advice Thread

[JimL]

10 hours ago, Ulfar4 said:

Hello everyone, I am going to build a DIY esl headphone but first I need some amplifier. I really do not want something fantastic, just to try out the headphones and then build something better. My inspiration was from John Broskie Tubecad schematic and head-fi forum. My question is will it play? And if not, what to change? Thanks for your help.

What I'm confused about is, why would you want to build a so-so amp just to try things out?  It seems to me that a better solution is to get an old Stax SRD-6 or 7 adapter box, modify it to pro bias and drive it with a standard speaker amp. That is a known quantity, will be easier, have at least decent quality sound, and probably less expensive.

[joehpj]

If low cost is the goal, JimL has a nice design SRX+. Simple and decent sound with fun of tubes. 

[Blueman2]

12 hours ago, JimL said:

What I'm confused about is, why would you want to build a so-so amp just to try things out?  It seems to me that a better solution is to get an old Stax SRD-6 or 7 adapter box, modify it to pro bias and drive it with a standard speaker amp. That is a known quantity, will be easier, have at least decent quality sound, and probably less expensive.

 

9 hours ago, joehpj said:

If low cost is the goal, JimL has a nice design SRX+. Simple and decent sound with fun of tubes. 

+1 on the above.  I started by buying an old SRD-7 off eBay and did the mod to give it 580V bias.  Nice basic project and it provides a good sound, if not exceptional.  Then, if you like that, move on to building the SRX Plus.  That amp is pretty simple to build and is a big step up from the SRD-7.  If I were doing this from scratch today, I would just go directly to the SRX Plus.  Great sounds and really fun learning experience.  

[Ulfar4]

16 hours ago, kevin gilmore said:

electrostat, dn2540 goes boom at anything over 400v. and its a depletion mode part, so not a great idea. what you are doing is a hev70 copy, look at the original for better ideas. also the opamps driving the mosfets have the feedback wrong.

I was assuming that the 600V supply will split to 300V per side. So I put in there 10m90s as css. Is my thoughts wrong then? I redraw the feedback little bit acording to hev70. If I am understanding correctly, the feedback is there for setting the offset? 

11 hours ago, JimL said:

What I'm confused about is, why would you want to build a so-so amp just to try things out?  It seems to me that a better solution is to get an old Stax SRD-6 or 7 adapter box, modify it to pro bias and drive it with a standard speaker amp. That is a known quantity, will be easier, have at least decent quality sound, and probably less expensive.

Yes, it will be easier. But for me, it's a lot more expensive than building something like that. Besides transistors I have all parts so it will be inexpensive.

7 hours ago, joehpj said:

If low cost is the goal, JimL has a nice design SRX+. Simple and decent sound with fun of tubes. 

I will sure look at it, but I wanted to avoid tubes. But I believe that it can sound very well. This is only for fun so I could experiment with headphones.

[mypasswordis]

[kevin gilmore]

the second schematic is not going to work, there is nothing to keep the output at +300, the left side of R4 needs to go to some negative value

and AGAIN, the dn2540 will BLOW UP when the voltage across it hits 400v

use a bsp125 in the to220 package if you want a 600v power supply

but much better is a c2m1000 and a 900v power supply

 

 

Edited November 3, 2017 by kevin gilmore

[JimL]

2 hours ago, Ulfar4 said:

I was assuming that the 600V supply will split to 300V per side. So I put in there 10m90s as css. Is my thoughts wrong then? I redraw the feedback little bit acording to hev70. If I am understanding correctly, the feedback is there for setting the offset? 

So, the problem is, even if the DN2540 sits at +300 volts, with signal, it can swing anywhere between ground and +600 volts with a big enough input signal.  But the DN2540 is only rated at 400 volts.  You should be able to see where I am going with this.  Furthermore, as Kevin Gilmore points out, there is nothing in the circuit that sets the drain voltage of the DN2540 at +300 volts.  The 10M90S current source is pretty much indifferent as to what the drain-source voltage of the DN2540 is so there is really nothing to control the quiescent voltage of the DN2540 drain.  The feedback resistor really won't control it.  I saw this in the SRX Plus, which uses 250k feedback resistors - there were times during de-bugging and adjustment when one output was nearly at B+ whereas the other was close to B- until I got things tuned up.  That is tolerable with tubes because the maximum tolerated voltage is higher than the maximum rated plate voltage.  Not so with MOSFETs or transistors - exceed the maximum for even an instant and you have a fried device.  If that happens to your circuit on turn-on, it'll blow so fast you won't even know what happened.

 

Take a closer look at the HEV70 schematic, it looks like there is a resistor network to set the quiescent output voltage.

Edited November 3, 2017 by JimL

[Ulfar4]

I was trying to fix it. I put negative voltage on the feedback as you said. I cannot find bsp125 as to220, but i found irfbe20 which is up to 800V. I hope this time I have done it right. Only problem is that I cannot put css in there without ruining output. With those 50k resistors the simulation worked. Anyway, how much voltage swing is actualy needed for esl headphones? Something like stax sr207 for expample.

[kevin gilmore]

that cannot work with the + inputs of the opamps grounded.

[Ulfar4]

OK, so I have to put there resistor like this?

[kevin gilmore]

like this, one side, replace r7,r8 with 10k pot

 

electroopamp.PDF

[Ulfar4]

Thanks, it helped me alot.

[mwl168]

I was PM'ing with another headcaser about the practice of lifting the resistors off the PCB board for high voltage applications. I improvised a method that I have used for all my builds so far and thought I would share it here in case others may be interested. 

See the attached pics, I used a lead bender I bought which has a provision that works perfect for most of Kevin's layout where lead spacing for the resistors is about 11mm.

To lift the resistors off the board, I put 5mm tall standoffs with a washer to make the total height about 5.5mm. This works great with KOA Speer 1/2W resistors which has a diameter of 3.5mm, thereby  provides a 2mm gap between the resistor and the board. I make sure I do this on a flat, hard surface so all the resistors come out uniform height. I usually solder one lead first , exam to make sure it was not crooked, then solder the other lead.  

Edited November 5, 2017 by mwl168

[luvdunhill]

Or just put a zip tie under and solder the top side of the board.

[makmeksam]

 

Today I found above information regarding Schiit Jotunheim. I was researching to buy an amp from Schiit. Are Schiit amps poor in design and performance? Any idea about Mjolnir 2?

[johnwmclean]

Epic!

[Blueman2]

Question about filament voltages.  I have seen numerous recommendations to slightly reduce filament voltage using an in-line 10W resistor of 0.5 Ohms or so.  Looking at my Antek transformer, its filament windings are rated at 6.3VAC at 115VAC input.  Here in SFO Bay Area, I average about 125VAC on my lines.    This gives me about 6.8V on the filaments, so almost 10% high.  I am not too concerned about lifespan of the tubes, so long as I get around 1000-2000 hours out of the output tubes (6SN7 in my case).  Is it recommended I add a voltage reducer?  For a pair of 6SN7s, would about 1 ohm be too much?  I am not worried about the input tubes (12AT7s) as much (or should I be?).

[mwl168]

30 minutes ago, Blueman2 said:

Question about filament voltages.  I have seen numerous recommendations to slightly reduce filament voltage using an in-line 10W resistor of 0.5 Ohms or so.  Looking at my Antek transformer, its filament windings are rated at 6.3VAC at 115VAC input.  Here in SFO Bay Area, I average about 125VAC on my lines.    This gives me about 6.8V on the filaments, so almost 10% high.  I am not too concerned about lifespan of the tubes, so long as I get around 1000-2000 hours out of the output tubes (6SN7 in my case).  Is it recommended I add a voltage reducer?  For a pair of 6SN7s, would about 1 ohm be too much?  I am not worried about the input tubes (12AT7s) as much (or should I be?).

I tried this before with my universal PSU for Blue Hawaii, Grounded Grid, etc. running on Antek transformers. My empirical experience is that a few tenth of volts  is lost (measured at the filament connectors on the amp PCB) from the transformer through the umbilical, internal wiring and the PCB trace. I took out the "gizmo" I rigged together to lower the EL34 filament supply inside the PSU.

My suggestion; don't do anything additional until you measure it on a running amp. 

[n_maher]

2 hours ago, Blueman2 said:

Question about filament voltages.  I have seen numerous recommendations to slightly reduce filament voltage using an in-line 10W resistor of 0.5 Ohms or so.  Looking at my Antek transformer, its filament windings are rated at 6.3VAC at 115VAC input.  Here in SFO Bay Area, I average about 125VAC on my lines.    This gives me about 6.8V on the filaments, so almost 10% high.  I am not too concerned about lifespan of the tubes, so long as I get around 1000-2000 hours out of the output tubes (6SN7 in my case).  Is it recommended I add a voltage reducer?  For a pair of 6SN7s, would about 1 ohm be too much?  I am not worried about the input tubes (12AT7s) as much (or should I be?).

Why not just regulate the line voltage before the amp? 

https://www.amazon.com/APC-LE1200-Automatic-Voltage-Regulator/dp/B00009RA60

Not endorsing that particular solution, just saying that there are cheap options that might be worth exploring.  

[JimL]

On 11/11/2017 at 12:22 PM, Blueman2 said:

Question about filament voltages.  I have seen numerous recommendations to slightly reduce filament voltage using an in-line 10W resistor of 0.5 Ohms or so.  Looking at my Antek transformer, its filament windings are rated at 6.3VAC at 115VAC input.  Here in SFO Bay Area, I average about 125VAC on my lines.    This gives me about 6.8V on the filaments, so almost 10% high.  I am not too concerned about lifespan of the tubes, so long as I get around 1000-2000 hours out of the output tubes (6SN7 in my case).  Is it recommended I add a voltage reducer?  For a pair of 6SN7s, would about 1 ohm be too much?  I am not worried about the input tubes (12AT7s) as much (or should I be?).

So, it was shown decades ago that 5-10% high filament voltages will decrease tube life.  On the other hand, running your filaments 5=10% low will not affect cathode emissions much.  6SN7 tubes draw 0.6 amps each, or 1.2 amps for a pair, so an 0.5 ohm resistor will drop your filament voltage to about 6.2 volts, while a 1 ohm resistor will drop it to 5.6 volts.  Somewhere around 0.7-0.8 ohms should do it - a wire wound resistor of those values is pretty inexpensive.

[luvdunhill]

Man I hate casework. I cut this one just a little two close.

[Pars]

^ I hear ya.

Was struggling with trying to decide whether this would work or not. Voltage switch (115/230) with 2 transformers, and switching of the main transformer on with front power switch (e24 type functionality, but allowing line voltage selection). Relays switching both primaries of the main transformer since both would be live in the case of 115 parallel connection. Maybe I'll just use a Meanwell AC-DC brick for the relay board power (ala Soren).

 

 

Edited November 18, 2017 by Pars

[spritzer]

I'm working on something similar with the T2 redesign.  I think I'll just end up using two switches as the LV stuff runs the delay relay. 

[luvdunhill]

$ python -m visa shell

Welcome to the VISA shell. Type help or ? to list commands.

(visa) list
( 0) GPIB0::16::INSTR
( 1) GPIB0::28::INSTR
( 2) GPIB0::1::INSTR
( 3) ASRL2::INSTR
( 4) ASRL1::INSTR
(visa) open 0
GPIB0::16::INSTR has been opened.
You can talk to the device using "write", "read" or "query.
The default end of message is added to each message
(open) query *IDN?
Response: KEITHLEY INSTRUMENTS INC.,MODEL 2015,0788786,B09  /A02  

(open) close
The resource has been closed.
(visa) open 2
GPIB0::1::INSTR has been opened.
You can talk to the device using "write", "read" or "query.
The default end of message is added to each message
(open) query *IDN?
Response: TEKTRONIX,TDS 3032C,0,CF:91.1CT FV:v4.25 TDS3GV:v1.00 TDS3FFT:v1.00 TDS3TRG:v1.00

(open) close
The resource has been closed.
(visa) 

Played around with PyVISA today...

[jose]

Novice question: 

What is the correct way to tin this type of pad? Should I fill with tin or just put a little under the component as I normally do?