Blue Hawaii BJT Build thread

[joehpj]

Here are my BOM for a pair of BH.

https://www.mouser.tw/ProjectManager/ProjectDetail.aspx?AccessID=b44788bc7f

for GRLV and GRHV. copied from Carbon build thread.

http://www.mouser.com/ProjectManager/ProjectDetail.aspx?AccessID=b6d687c7f7

http://www.mouser.com/ProjectManager/ProjectDetail.aspx?AccessID=54687aa8fa

Parts I already have or going to buy from local dealer are NOT included like LEDs, insulators, IC sockets. Someone need to check if all resistor value are correct.

I hope I could finish it at the end of April 

[Pars]

Probably should ask this in the GRLV thread, but I'm curious why the BOMs I see show Elna Silmic caps for the 220uf caps on the PSU? Usually audio grade caps are spec'd for coupling caps (Pass seems to like them), but for a PSU? Some reason that I am missing? I used Pana FM/FR for the ones I built.

 

 

Sent from my iPhone using Tapatalk

 

[livewire]

Quote

 

BH BJT BOM MOUSER part number 80-PHE426PB6100JR06 references a .1uF /1kv cap.

Is this a typo? (should be a 1uF / 450v cap as shown on the BH pcb silkscreen?)

 

Edited March 28, 2017 by livewire

[JoaMat]

The 33pF and 220pF need to be at a higher voltage? I use 1000 V Cornell Dubilier micas.

[joehpj]

16 hours ago, livewire said:

BH BJT BOM MOUSER part number 80-PHE426PB6100JR06 references a .1uF /1kv cap.

Is this a typo? (should be a 1uF / 450v cap as shown on the BH pcb silkscreen?)

 

I copied from carbon BOM. Those are bypass cap so use whatever you want/have.

1 hour ago, JoaMat said:

The 33pF and 220pF need to be at a higher voltage? I use 1000 V Cornell Dubilier micas.

My bad. should at least >400V to be safe.

[livewire]

Does anyone have a schematic for this build that they could post here?

 

[mwl168]

On 4/11/2017 at 12:07 PM, livewire said:

Does anyone have a schematic for this build that they could post here?

 

I don't have a complete updated schematic but the hand-draw images below are the changes I traced on the updated Blue Hawaii. IIRC, the rest of the circuit is the same as the original one (the color image). I built the 2SK216/2SJ79 version so replace the 2SJ79 in the drawing  with KSA1220 KSC2690 . The KSA1156 emitter resistors determine the current to the third stage and the current is reduced in this version.

I will add some notes from my own build later.

  

Edited December 24, 2017 by mwl168

[mwl168]

some build notes:

·         The 360R KSA1156 emitter resistors set the current to the 3^rd stage. The original BH ran 10mA, the 360R resistors in current version set it to be about 3mA to reduce heat.  If you decide to raise the current, make sure you compensate for it by lower initial value in offset adjustment.

·         The boards are designed T2 style – the tubes are mounted on the back side of the board. I also mounted the LED, all trim pots and the 50R CCS resistors on the back side for ease of diagnosis and adjustment.

·         Be VERY CAREFUL when you measure the voltage across the 50R CCA resistor next to the 24V zener. There is significant voltage potential difference between the two and they can get arcy sparky easily – it happened to me twice when I was testing/adjusting the CCS current - big bang, magic smoke and flying silicon pieces. I ended up mounting those 50R resistors tall to give additional Clarence.

·         Also be careful with the DN2540 in the CCS, their exposed tabs are live.

·         It should be obvious but the PZTA06 and PZTA56 SMD are to be mounted vertically observing the pinout.   

[jose]

Thanks for advise mwl168.

[livewire]

@mwl168 - Thanks much for posting the schematics!

I have a question, what's up with the STN9360 BJTs?

Are these an alternate for the other pair of KSA1156 shown?

(to be used INSTEAD OF the KSA1156 that are between the tube socket traces)

I ask because I've seen a lot of EOL parts spec'd before

and SMT seems to be taking over. 

[mwl168]

2 hours ago, livewire said:

I have a question, what's up with the STN9360 BJTs?

Are these an alternate for the other pair of KSA1156 shown?

(to be used INSTEAD OF the KSA1156 that are between the tube socket traces)

 

That's correct. You either use the STN9360 or the KSA1156 but not both.

I believe both are current production parts available from Mouser. I agree that SMT is taking over. I wonder how much longer through-hole components will continue to be made?

 

[JimL]

6 hours ago, mwl168 said:

some build notes:

·         Be VERY CAREFUL when you measure the voltage across the 50R CCA resistor next to the 24V zener. There is significant voltage potential difference between the two and they can get arcy sparky easily – it happened to me twice when I was testing/adjusting the CCS current - big bang, magic smoke and flying silicon pieces. I ended up mounting those 50R resistors tall to give additional Clarence.

·         Also be careful with the DN2540 in the CCS, their exposed tabs are live. 

One suggestion, if you're setting up the current for the output current sources, you can do this using a low voltage source, e.g. 2 9V batteries in series across the current source for preliminary set-up.  Because of their high impedance, the current will change very little between low and high voltage.  Less chance of arcy sparky.

[Laowei]

Any opinions on the advantages of revisiting using BJTs as current sources (as in the original KGBH), as opposed to the latest FET cascode design?

Is there an increase in performance. or just another way to skin the cat with currently available parts? 

Edited April 19, 2017 by Laowei

[Whitigir]

What does BHSE use for delay mechanism ? Perhap a Separated switch is better than timer relay board ?

Edited April 19, 2017 by Whitigir

[sorenb]

2 hours ago, Laowei said:

Any opinions on the advantages of revisiting using BJTs as current sources (as in the original KGBH), as opposed to the latest FET cascode design?

Is there an increase in performance. or just another way to skin the cat with currently available parts? 

Cascoded bjt's CSS are linear throughout the audio region.
In comparison the cascoded fet's demonstrate impressive results at DC but drops through out the audio region, and 'meets' the cascoded fet CSS around 3kHz..

I use a T2 CSS in my Blue Hawaii. Don't use the T2 CSS in my Carbon anymore.

[Laowei]

Thanks Soren. 

[JimL]

On 4/19/2017 at 2:30 PM, sorenb said:

Cascoded bjt's CSS are linear throughout the audio region.
In comparison the cascoded fet's demonstrate impressive results at DC but drops through out the audio region, and 'meets' the cascoded fet CSS around 3kHz..

I use a T2 CSS in my Blue Hawaii. Don't use the T2 CSS in my Carbon anymore.

I'm not sure that is the case.  On the Gary Pimm website (which seems to have disappeared), he gave measurements of cascoded FET CCS which showed performance flat over and above the audio range - in fact, as good as a very high impedance resistor.  Same for Walt Jung's articles, which showed very high, flat impedance to well above 20 kHz.  And that makes sense, because although the disadvantage of MOSFETs is high capacitance, in the cascode CCS the "lower" MOSFET sets the current, and the "upper" MOSFET largely shields the lower MOSFET from voltage variations, so that it sees a nearly constant voltage over a large voltage swing, and therefore produces a constant current regardless of voltage swings, which translates to a high and consistent impedance even to very high frequencies, equivalent to a very low capacitance.

[sorenb]

I am not aquinted with the work of Gary Pimm, and as you stated, his website is no longer up.

10 hours ago, JimL said:

Same for Walt Jung's articles, which showed very high, flat impedance to well above 20 kHz

Not really, Jungs measurement shows deterioation above 1kHz - Jung even mention this in his text.

Still, the cascoded fets demonstrate higher impedance in comparison to their bjt counterparts. - but seems to lack in linearity.

[kevin gilmore]

the closest thing to a perfect current source is a high voltage tube with a cathode resistor, i.e. megatron

T2 current source is actually not really very good because the resistor divider for the pnp current source causes a significant change in current

for example 700v peak to peak on the plate the current varies from 13.3 to 15.8ma

same thing for a cascoded depletion mode is 14.8ma to 14.85ma

a better idea would be to replace the stacked pnp and resistor with a very highly isolated dc to dc converter, something good for 2kv of isolation and you would need 4 of them per amplifier.

Edited April 24, 2017 by kevin gilmore

[joehpj]

But mid point of megatron seems to drift quiet a bit. Is that because of the tube ccs?

[kevin gilmore]

its ac coupled and there is no servo

[spritzer]

You will always get drift with tubes, just the nature of the beast. 

[kevin gilmore]

here are 3 current sources at 10khz

 

currentsource.PDF

Edited April 24, 2017 by kevin gilmore

[mypasswordis]

Uh oh, I see GeorgeP is crazy enough to want to set up another T2 run... are there going to be some CCS modifications to the board?  

[Laowei]

FYI,

Gary Pimm's pages have been deleted for some time. Here is a link to the archived pages on The Wayback Machine. Unfortunately. the CCS measurement page did not make it. The other links work OK. 

http://web.archive.org/web/20151117131328/http://pimmlabs.com/

I had a few screen shots of Gary's CCS impedance measurements saved: 

Edited April 24, 2017 by Laowei