[BoilermakerFan]

Quote:

I decided to go with the 6N27P.  Have four of them on the way to me.

Regular or EV versions?
 
Edit:  Nevermind, thought it was the 6n23P...

 

[stringgz301]

I remember reading earlier in this thread that in this updated design device matching was not as important. That being said, which devices should be matched?

 

[runeight]

MHO is that device matching is not needed. The Toshiba transistors should be selected in the same hfe class though.
 
If I were matching I would match two things.
 
1. The O/P mosfets,
2. The mu of the two triode sections or even all four sections in both channels.
 
 But if I were building the amp I would not make the effort to do this. I do understand that others have different diy preferences.

 
 

 

[luvdunhill]

If I remember correctly, these MOSFETs (IRF550/IRF9540) will not make great complementary pairs... I bet you'll have a ~0.5V difference in Vgs between the two types. If you have a distortion analyzer, you could probably play with R22, R23 to find a null point... probably not worth the effort as Alex said though.

 

[sachu]

Indeed, matching of the triode sections of the tube might actually be more beneficial, even though you can balance these out (so far as they are not too far away from each other). 
 
Got my Toshiba BJTs at the ready for the output devices.

 

[luvdunhill]

yup, however if you're not matching either at the voltage/current/temperature parameters of the circuit, it might be wasted effort that could be spent in building another one of Alex's designs

 

[runeight]

Gents, please don't use the BoM that I just posted. I used an older version and it is very wrong.
 
New one up soon.

 

[runeight]

New BoM is posted. I believe this one is correct. Please let me know.

 

[BoilermakerFan]

Quote:

yup, however if you're not matching either at the voltage/current/temperature parameters of the circuit, it might be wasted effort that could be spent in building another one of Alex's designs

Muaaahaaahahaaa! 

 
But on the DIY front for dynamic cans, the EHHA is the top of the heap with the possible exception of a well executed Bijou if you prefer it's sound. There's more to be eeked out of the EHHA WRT to output sand rolling too, then you have to tweak the NFB.  I won't be matching my devices, but I will be rolling them. 

 

[runeight]

No boards today. Looks like the deluge has slowed down the fedex deliveries. Oh well, no one could have gotten a board before the weekend anyway. Maybe tomorrow.

 

[runeight]

Wrong again. Even thought the tracking said that the package was still at the facility, the boards have arrived. Very clever boards gents.

 

[jdkJake]

yup, however if you're not matching either at the voltage/current/temperature parameters of the circuit, it might be wasted effort that could be spent in building another one of Alex's designs

Ouch!



BTW, you are right, there is up to a 1/2 volt difference between the two parts. At least at the sub-optimum level I tested them at.

 

[ujamerstand]

Quote:

MHO is that device matching is not needed. The Toshiba transistors should be selected in the same hfe class though.
 

Doh. The sources of my 2sa1145s are different, some from bdent, some from mouser. In fact, they are visibly different. -_-\\ I should probably get a test rig going to check their hfe value before I populate them into the boards...

 

[jdkJake]

Doh. The sources of my 2sa1145s are different, some from bdent, some from mouser. In fact, they are visibly different. -_-\\ I should probably get a test rig going to check their hfe value before I populate them into the boards...

Toshiba uses two ratings, O and Y. The O rating is the lower of the two (80 to 160 hfe) with the Y being the higher (120 to 240). There is the potential for overlap, but, I think Alex is trying to say that as long as you stay within the same rating class you are fine.

 

[luvdunhill]

Quote:

Doh. The sources of my 2sa1145s are different, some from bdent, some from mouser. In fact, they are visibly different. -_-\\ I should probably get a test rig going to check their hfe value before I populate them into the boards...

even if they are visible different, that's okay as long as the hfe class is the same. There are two versions of these, one with etched identifiers on the case, and one with whiteish-ink-printed labels. The etched ones are the later devices.