Technical Assistance/Advice Thread

[n_maher]

Since it appears to be a problem on the input side, maybe try some RCA/XLR cables or adapters so that you can bypass the potentially faulty RCA jack?  

[kevin gilmore]

it is possible that parts of the power supply are dual mono and one side is getting too hot for some reason, or one of the electrolytics is causing trouble. Get an infrared thermometer and see if anything is getting hot.

[dncnexus]

I bought both an RCA/XLR cable and a infrared thermometer coming in today/tomorrow to test out. I will update with information when I test.

[spritzer]

I thought some electrolytics looked off but can't tell with any certainty from the pics.  Make sure none of them are domed or have any leakage around them. 

[dncnexus]

3 hours ago, spritzer said:

I thought some electrolytics looked off but can't tell with any certainty from the pics.  Make sure none of them are domed or have any leakage around them. 

Just checked all the electrolytics and none seem to be domed or leaking at all. 

[kevin gilmore]

i'm assuming that all the meter numbers don't change when the channel cuts out? which kind of eliminates the front end tubes and circuitry?

 

[dncnexus]

2 minutes ago, kevin gilmore said:

i'm assuming that all the meter numbers don't change when the channel cuts out? which kind of eliminates the front end tubes and circuitry?

 

The meter numbers only for the bias of the EL84 tubes, but no I don't notice any number differences. I can check again just to verify 100%

[kevin gilmore]

know anyone you can borrow a oscilliscope from. (with high voltage 100x probe) ?

 

[dncnexus]

3 minutes ago, kevin gilmore said:

know anyone you can borrow a oscilliscope from. (with high voltage 100x probe) ?

 

I don't. Can I just buy a regular oscilliscope and source a HV100x probe for it? I can probably try to source one

[kevin gilmore]

a scope is a good thing to have. plenty of high voltage probes on ebay.

like this

Oscilloscope Scope Test Probe P4100 100:1 High Voltage Withstand 2KV 100MHz 1.2M | eBay

Edited November 2, 2021 by kevin gilmore

[dncnexus]

On 11/2/2021 at 3:52 PM, kevin gilmore said:

a scope is a good thing to have. plenty of high voltage probes on ebay.

like this

Oscilloscope Scope Test Probe P4100 100:1 High Voltage Withstand 2KV 100MHz 1.2M | eBay

I have an oscilloscope and that test probe you linked coming. I have a SE to balanced cable arriving today and can try the BAL inputs to see if maybe can isolate the problem area.

[kevin gilmore]

Best place to measure is the scope set to AC and tied to either side of the janzen caps.  the side next to the gain tubes is likely to be about +150v dc with music riding on it, probably 15 volts of peak to peak AC when volume is cranked.

There should be music on both sides of both janzen caps.

If there is still music on the caps when the sound disapears then the problem is the output tubes or output transformer

 

[dncnexus]

Update:

Received the RCA->XLR cable and tested it out with BAL inputs. Problem persists, so it must be somewhere outside of the inputs. The Oscilloscope should be arriving in a week/week and half, will update with more information then/if I find test anything else out.

Thanks all for the assistance so far.

[n_maher]

If you can, inspect the solder joints at the tube sockets.  I had a power amp develop a cold joint at one and it cause enough of a problem that it cooked a trace on the PCB and pretty much rendered the unit a total loss.  

[kevin gilmore]

the janzen caps are definitely between stage 2 and stage 3.  The orange caps are the output caps.

the output is a center tapped inductor, not a transformer. B+ is 525 volts on the center of the transformer.

Edited November 5, 2021 by kevin gilmore

[ibuski]

Hello everyone!

I thought I should build Dynahi from the parts I purchased more than 15 years ago.  Yes, original Dynahi.  I was kept busy in decade for job or for family.  Now my daughters are independent and I work less.   I should use my free time in something productive.   Before starting amplifier building, I made up a Tr/JFET measurement device and started measuring those devices.   Since I had so many transistors, Tr matching wasn't difficult.  An issue is JFET, K389 and J109, since I have only 4 pairs.   Luckily all JFETs have good match within a package, so I'm talking about matching between N ch and P ch.

I read old HeadFi thread about device matching, and found that people often measure only Idss for FET.   I understood that the FET's characteristic is defined by both Idss and Vgs.  They are usually correlated, but not well, right?   So I measured both Idss @ 10 V and Vgs @ 1mA because I read that 1mA is working condition of them in Dynahi.  Unfortunately I can't find even one pair with close enough for both parameters.   Some have very close Vgs such as 290 vs 300 mV, but Idss showed more than 2 mA difference.  One pair has good match in Idss, but Vgs are 390 vs 290 mV, totally different.  My question is which parameter should I prioritize for Dynahi input stage usage?  I tried to calculate the slope in V/A, but all K389 have ca 45 V/A of slope while J109 have ca. 35 V/A and don't fit.   I'm tempted to match Vgs and allow Idss gap of 2 mA...   Any advise?

[federicopol]

Hello everyone, here asking for some help wuth a build i am planning:

I'd like to build a amplifier that has:

pro bias and normal bias.

Sounds reasonably good,

Proven circuit.

Simple to make and not to much expensive because i want to build up skills and confidence before a more serious build.

So i choose the tubecad curcuit (stealing the schematics plus PS that @chinsettawong posted on his thread on the other forum) the only problem is that i also need normal bias.

 

I plan to put a voltage divider on the BIAS PS and take 230v from that. Is my curcuit modification ok? Did i commit basic newbie errors that will kill me and burn my house?

Or is the voltage divider a stupid idea per se?

i plan to use an hammomd rrafo for valve amplifiers for HV and Heater and two small trafos with one wired backwards for the bias.

Thank you very much

 

[spritzer]

There are far simpler ways of making that bias supply, a simple voltage divider off the B+ would work just fine. 

For the pro bias, I'd put an adjustable voltage doubler off the main transformer, see something like the original KGSS article on Headwise memorial. 

[federicopol]

4 minutes ago, spritzer said:

There are far simpler ways of making that bias supply, a simple voltage divider off the B+ would work just fine. 

For the pro bias, I'd put an adjustable voltage doubler off the main transformer, see something like the original KGSS article on Headwise memorial. 

Thank you very much.

That way i will need one trafo only at expense of some more small parts.

Will study, will copy, then probably ask again if doing it correctly.

 

Thank you very much again

[Sathimas]

Hi there, 

I have a kind of annoying problem.

I got teflon stax jacks from the group buy here.

They fit perfectly with m old Lambda Pro but are too tight for the 007MkI.

To pull the plug out, I have to apply force and secure the amp with the other hand.

Is there any way to widen them?

Cheers

Sathimas

 

 

[kevin gilmore]

take them completely apart and use male xlr pins to open up the socket pins

[ibuski]

On 11/18/2021 at 10:07 PM, ibuski said:

Hello everyone!

I thought I should build Dynahi from the parts I purchased more than 15 years ago.  Yes, original Dynahi.  I was kept busy in decade for job or for family.  Now my daughters are independent and I work less.   I should use my free time in something productive.   Before starting amplifier building, I made up a Tr/JFET measurement device and started measuring those devices.   Since I had so many transistors, Tr matching wasn't difficult.  An issue is JFET, K389 and J109, since I have only 4 pairs.   Luckily all JFETs have good match within a package, so I'm talking about matching between N ch and P ch.

I read old HeadFi thread about device matching, and found that people often measure only Idss for FET.   I understood that the FET's characteristic is defined by both Idss and Vgs.  They are usually correlated, but not well, right?   So I measured both Idss @ 10 V and Vgs @ 1mA because I read that 1mA is working condition of them in Dynahi.  Unfortunately I can't find even one pair with close enough for both parameters.   Some have very close Vgs such as 290 vs 300 mV, but Idss showed more than 2 mA difference.  One pair has good match in Idss, but Vgs are 390 vs 290 mV, totally different.  My question is which parameter should I prioritize for Dynahi input stage usage?  I tried to calculate the slope in V/A, but all K389 have ca 45 V/A of slope while J109 have ca. 35 V/A and don't fit.   I'm tempted to match Vgs and allow Idss gap of 2 mA...   Any advise?

I ended up relying on matched quad sell on eBay, famous punkydawgs.  Of course he sells genuine Toshiba JFET.  I bought 2 sets.  It took ca. 100 EUR incl. shipping to Germany.   The first set was perfect!   J74: 247mV@1mA, IDSS 7.8mA, 229mV@1mA, IDSS 7.8mA, K170: 268mV@1mA, IDSS 7.8 mA, 270mV@1mA, IDSS 7.8mA.   The second one contained one J74 with 0.9mA higher IDSS although all other three stayed at 8.2mA.   He advertises that all his FETs were tested at 72 degree, probably it's 72 degF so this means 22 degC, slightly higher than my room, but I wonder if that can cause such error although all others were so accurate.

OK, now I got perfect one quad set.   Probably it's time to give up and find best compromise from the rest of parts for the other channel.

[Pars]

Those in the other set will be fine.

[Craig Sawyers]

The alternative is to buy Linear Systems LSK170/LSJ74 matched pairs, quads and even octets from DIYAudio's store.

[ibuski]

6 hours ago, ibuski said:

I ended up relying on matched quad sell on eBay, famous punkydawgs.  Of course he sells genuine Toshiba JFET.  I bought 2 sets.  It took ca. 100 EUR incl. shipping to Germany.   The first set was perfect!   J74: 247mV@1mA, IDSS 7.8mA, 229mV@1mA, IDSS 7.8mA, K170: 268mV@1mA, IDSS 7.8 mA, 270mV@1mA, IDSS 7.8mA.   The second one contained one J74 with 0.9mA higher IDSS although all other three stayed at 8.2mA.   He advertises that all his FETs were tested at 72 degree, probably it's 72 degF so this means 22 degC, slightly higher than my room, but I wonder if that can cause such error although all others were so accurate.

OK, now I got perfect one quad set.   Probably it's time to give up and find best compromise from the rest of parts for the other channel.

Surprisingly, punkydawgs gave me full refund immediately when I reported my measurement result to him.   He wrote he guarantees 0.1 mA range and apologized for the error.  I expected argument, but he just trusted me.   What an honest seller he is!   His pricing isn't low, but I learned importance to buy from reliable seller.

Thanks Craig for the information about DIYAudio.  I'm a member of it, but wasn't aware of their store contents.  Now I can consider that also as one source with refunded budget.

Pars, yes, they aren't bad at all.  That's still less than 1 mA of range.   Maybe I'm unnecessarily picky.   I don't know how it affects amplifier's performance.   Probably my ears can't recognize that influence.