Blue Hawaii BJT Build thread

[sorenb]

45 minutes ago, Blueman2 said:

1) Is it necessary or common to have soft stop on the BH?  I am hoping I do not need this.

Not needed.
Soft START is a good thing to be easy on the tubes. T2 uses 60 seconds delay.
Some uses constant current sources to feed the filament, to avoid the high current when the tube is cold  - high current is stressing the tube (and usually why incandescent bulbs fail)... some runs the filament a bit below spec e.g. 6V rather than 6.3V,  (even seen 5.5V) as far as I have measured 6V is pretty safe, where as lower voltage increase the risk for "cathode stripping" 

[Blueman2]

Thanks sorenb!  I will just stick with soft on and adjust to 60 seconds.  And since my mains run about 125V here (vs 115 spec for my trafo) I will add a power resistor to the 6.3V to get that down to 6V.  ??

[sorenb]

calculate the resistors based on measured values when everything are hooked up, and tubes are running warm - depending on the transformer the spec'ed voltage might be different under load. 

[kevin gilmore]

Soft start is one thing, waiting 60 seconds to turn on the high voltage is something else and is a very good idea

[Blueman2]

44 minutes ago, kevin gilmore said:

Soft start is one thing, waiting 60 seconds to turn on the high voltage is something else and is a very good idea

Ah, I had my terminology wrong then.  So soft start refers more to limiting the current going into the filaments (or others parts if desired?) to lengthen tube life?  Whereas the delayed HV is another separate option in order to avoid the offset spike?  

Edited January 9, 2018 by Blueman2

[kevin gilmore]

although you certainly can (singlepower es1,es2...) when you apply high voltage with the tube cold, the cathode can strip. with the diy-T2 applying high voltage with the tubes cold does all sorts of bad to the input tube, putting +500 on the cathode with the filament grounded.

soft start on an el34 seems like a waste of time, haven't seen anyone do that.  On expensive DHT's, might be a good idea, might be a waste of time.

[Blueman2]

On 1/6/2018 at 3:53 PM, mwl168 said:

I started out using 150R instead of the 360R for the KSA1156 emitter resistors to set the current to the 3^rd stage and never tried the 360R on my build so cannot comment on the sonic difference. I believe the original BH ran 10mA current to the 3rd stage and was later reduced to 3mA to reduce the heat. I decided to go half way based on a similar adjustment I made on my KGSSHV and in that case I like the sound of the raised current much better. 

Do note though that if you raise the current to the third stage you need to make compensation to the offset-setting resistor and trim pot. I would suggest using 1.2K or so trim pot instead of the 2K and preset it to 600R or so. You'll need to measure and do some calculation so you do not have too much voltage drop across the offset trim pot. 

I plan to use 200R resistor and see how that plays out.  Since I am using pretty large heatsinks (90mmx300mm, 50mm long fins) I think I should be able to handle 10-15ma no problems.  To measure current to the 3rd stage, do I just measure the voltage across the 200R resistor I will be using and divide by 200? 

I will measure the voltage across the offset pot to determine current flow and power (my pot is rated at 1/2W power and 100ma across the slider).     

[mwl168]

32 minutes ago, Blueman2 said:

 To measure current to the 3rd stage, do I just measure the voltage across the 200R resistor I will be using and divide by 200? 

Yes.

[Blueman2]

Thanks.  Since I can only go with 1K or 2K trim pots (no other ratings available on this package) I will leave the 2K in, measure the voltage across it, derate it given it's resistance %, and see if it is OK.  Assuming 500R setting on the 2K trimpot, that gives derated power of 1/8W (1/2W full rating * 500/2K).  That means I can have max of 8V across the trimpot if my math is correct.     At 750R setting, I can have about 11.8V for max ratings.  We shall see!

Edited January 11, 2018 by Blueman2

[mwl168]

In my case, I also kept the 2K trim pot for the offset. It's a messy job desoldering a three pin part. It has worked fine so far just a bit too sensitive to the adjustment than I like. 

I don't remember the schematic much but if needed, you can always change the fixed resistor (if there is one) in series with the trim pot like in the KGSSHV.  

[Blueman2]

No fixed resistor; just the trim pot.  Good to hear the 2K is still running for you.  Since I will be a bit higher than you in my  KSA1156 emitter resistors (200R vs your 150R), thus lower in current to the 3rd stage, I should be OK.  I will report back once I have this powered up, in about a week I think.  

Edited January 11, 2018 by Blueman2

[Blueman2]

For the silicon that connects to the offboard heatsinks, are the 10M90S's the only ones that require aluminum oxide pads and non-metallic peek screws?  Everything else appears to be fully encased.  The 8N80C even appears encased and might not need the aluminum oxide pad and peek screws (?).   I wanted to confirm before I soldered those on.....

EDIT: Here is what I have done for pads.  Al oxide for the 10m90s and silicone rubber for the rest. 

Edited January 13, 2018 by Blueman2

[jose]

I used aluminum oxide pad for all. I don´t know if it´s necessary but I feel more secure. 

[Blueman2]

For the delay start of HV on the GRHV boards, what value resistor should I use for driving the CPC1117N Opto using 15V?  The spec sheet was unclear on this, but I was assuming I should add a 3K or so resistor (?).   

 

EDIT: Well, to answer my own question, I was able to find the forward voltage drop of the diode in the CPC1117N using my meter.  It is 1.2V.  So with 15V supply, that leaves voltage of 13.8V to get rid of.  The turn on amps for the 1117N is 1ma min to 50ma max.  Assuming 10ma as a safe target, I get R=13.8V/.010A=1.38K resistor.  Since 2.2K is a common value, I will go with that which will provide about 6ma current to trigger the opto relay. 

EDIT2: Yes, these values worked fine.  Here is a crude drawing of the circuit.  I am measuring about 5ma to the CPC1117N input, which is about what I wanted. A 3K resistor would probably work just fine as well and would lower the current even more to the CPC diode.  I am using THIS device, 12V version, from eBay as the timer.  It works great for this application and is cheap.  I have it set for delay on of 60s.  You can also put jumper on S5 on the device and it will provide the VCC without the additional wire from Vcc to relay COM.  

 

Edited January 20, 2018 by Blueman2

[Blueman2]

On 1/9/2018 at 10:56 AM, sorenb said:

calculate the resistors based on measured values when everything are hooked up, and tubes are running warm - depending on the transformer the spec'ed voltage might be different under load. 

 

I got the BH BJT up and running today.  Everything seems great and I was able to dial all the voltages and currents in just right.  One issue I am seeing is filament voltage.  My Antek PSU is giving me 6.8V under load for the EL34s.  Our voltage here is always around 125V, and the PSU was rated for 115, thus the higher value for the 6.3 feeds.  I put a pair of 5W 1Ω resistors in parallel to create a 0.5Ω resistor for each amp board filament supply.  That got the voltage down to 5.5VAC.  Is that too low or should I go out and buy another 1Ω resistor to add in parallel?      

[mwl168]

Congratulations!

I think the 5.5v is too low for the filament supply for the EL34. This is an often debated topic whether lower filament voltage is beneficial or not but I think most believe it should be within 10% of the 6.3v.

Edited January 18, 2018 by mwl168

[Blueman2]

Just now, mwl168 said:

I think the 5.5v is too low for the filament supply for the EL34. This is an often debated topic whether lower filament voltage is beneficial or not but I think most believe it should be within 10% of the 6.3v.

 

Hmm.  Then more like a 0.3Ω resistor is called for.  Or, if I am willing to live with a bit shorter tube life, is 6.7 to 6.8 VAC something I can live with without reducing?  This makes me realize that my 2 SRX-Plus amps are probably also being driven too high for filament voltage.  

[jose]

Other BH working !!

I think that 5,5Vac is low.

6Vac can be right. You can try with three 1R in parallel (about 0.33R) 

[sorenb]

13 minutes ago, Blueman2 said:

 

I got the BH BJT up and running today.  Everything seems great and I was able to dial all the voltages and currents in just right.  One issue I am seeing is filament voltage.  My Antek PSU is giving me 6.8V under load for the EL34s.  Our voltage here is always around 125V, and the PSU was rated for 115, thus the higher value for the 6.3 feeds.  I put a pair of 5W 1Ω resistors in parallel to create a 0.5Ω resistor for each amp board filament supply.  That got the voltage down to 5.5VAC.  Is that too low or should I go out and buy another 1Ω resistor to add in parallel?      

you need > 6.0V, 3x 1ohm might get you there, else use 4x
 

[Blueman2]

1 hour ago, mwl168 said:

... I think most believe it should be within 10% of the 6.3v.

 

1 hour ago, jose said:

....6Vac can be right. You can try with three 1R in parallel (about 0.33R) 

 

1 hour ago, sorenb said:

you need > 6.0V, 3x 1ohm might get you there, else use 4x

Thanks guys!  Just ordered a 10 pack of 0.3Ω 5W resistors and will get that voltage down to 6.0V - 6.3V range.  You see any harm leaving at 6.8V for now until I get the parts?  As mwl68 says, that is still within 10%.  Just barely.  

 

EDIT: I did find the spec sheet on the Tung Sol EL34Bs I am using.  It gives Filament range of 5.5-7.0.  So I guess I should be OK with limited use until my resistors arrive....

el34b-tungsol.pdf

Edited January 18, 2018 by Blueman2

[jose]

In your case, I would feel more comfortable using the amp at the minimum and without letting the valves get too hot.

[mwl168]

So the tubes you are using now are drawing 2.6A filament current per pair based on the measurements you provided. Since the filament supply is not regulated, subject to mains fluctuation and different tubes will likely varies a bit in their filament current draw. I suggest you make some calculation and set up the resistors to drop the filament supply under load to fall in a range you feel comfortable with.

I have some high cost NOS EL34 tubes so I personally feel most comfortable having the filament be between 6 - 6.3v.

And, by the way, I believe the filament current draw varies a bit from cold start to fully warmed up too.

[JimL]

I like to run my filament 5% low, which is 6 volts. I believe David Berning, who designs for long tube life, likes to run his 12.6 volt tubes at 11.5 volts, and his 6.3 volt tubes at 5.7 volts, which is close to 10% low.

[Blueman2]

I am moving the discussion from Carbon thread to here since it is more related to Blue Hawaii. 

Here is the prior discussion: 

 

I did not realize that the BH demands more current than the standard GRHV is set for.  There is a 5R1 resistor in the GRHV that needs to be changed to allow it to provide enough current for the BH BJT.  I am measuring current of 83ma on the B+ and 134ma on the B- on my build.  B+ rail is fine, but I had to reduce the current limiting resistor in B- from 5R1 to to 3R in order to get enough current to drive the BH.  Moving discussion here so that others can learn from my error.  

Edited January 20, 2018 by Blueman2

[sorenb]

17 minutes ago, Blueman2 said:

 ...But stays in a range of about 390 +/-3v.  Enough to allow me to get offsets down to under a volt and confirm the outputs to 20ma.  

assume you GRHV voltage out is 390V under load and +404 and -410 without? what is the voltage at the drain of the Sic (middle leg) against ground? 
and what is the voltage between R9/R7?  both with and without load