Replacing power MOSFETs in SDS Labs headphone amplifier

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Jump to Latest
P

pbilous

Member
Joined 2011
Paid Member
I need headphone amp and decided to go with schematics that uses OPAMP followed by push-pull transistors which provide current drive. Schematics that I found is SDS Labs Headphone Amplifier described on headwise.com (SDS Labs Headphone Amplifier | HeadWize). The schematics has PCB layout already developed for it. The AMB modified (AMB β18 Stereo Headphone Amplifier) it slightly like this:

beta18_ampsch.png


Original schematics is using Harris RFP15N05/RFP15P05 pair. The AMB version made with IRFZ24N/IRF9Z34N pair.

My idea is to replace output MOSFETs for IRF640/9640 or IRFP240/9240. The reason is because these are widely used MOSFETs that are available on eBay for moderate price.

Questions here are:
1. Will it be feasible to use such MOSFETs in the place of old ones?
2. If I use them, what changes, if any, should be made to the schematics?


Thank you
 
A

a1095us

Member
Joined 2006
I have built the AMB 18 in it's original form. It's a quality headphone amp. His original parts are not that inaccessible. I don't get your obsession! Why this form?
There are plenty of opamp push pull designs out there that will perform in a similar fashion. Even the Elliot Sound Products version is practically indistinguishable from this using fairly common parts.
 
P

pbilous

Member
Joined 2011
Paid Member
Thanks everyone for reply.

I guess I have to change the original question here to:
What are the benefits of using IRFZ over other MOSFETs? Why Ti Kan, author of schematics, used them?

And for the sake of experimenting, I am thinking of trying IRFZ AND other MOSFETs. If anyone has worked with following transistors, can you share experience working with them?
Transistors: IRF640/9640, IRF610/9610, IRF510/9510?
Or even these: 2SK1058/2SJ162, IRFP340/9240?

Will the sound change if I switch them? Or I will not be able to hear the difference at all?

Thanks again.
 
P

pbilous

Member
Joined 2011
Paid Member
a1095us said:
I have built the AMB 18 in it's original form. It's a quality headphone amp. His original parts are not that inaccessible. I don't get your obsession! Why this form?
There are plenty of opamp push pull designs out there that will perform in a similar fashion. Even the Elliot Sound Products version is practically indistinguishable from this using fairly common parts.
Click to expand...

For CCS did you use CR470 diode, or substitute? I could not find that part in stores. I was thinking about replacing it with LM317? Like this:
lm317t-current-regulation-circuit.gif


with 266 Ohm resistor for 4.7mA. Website - LM317 Current Calculator - Electric Circuit

Thank you.
 
P

pbilous

Member
Joined 2011
Paid Member
Thanks again everyone for replies to my dumb questions :)

I ended up buying IRFZ like in original schematics.

On AMB website (http://www.amb.org/audio/mmm/) I found tons of information on this schematics, concept of which they use to build M3 headphone amp. I also found an answer to my post to a1095us: AMB uses opamp bias that controlled by JFET cascode current source like in the schematics below.

Among other findings: head-fi forum thread (M³ Project Announcement) where they discuss M3 output stage MOSFET selection and reasons why to use IRFZ24N/IRF9Z34N pair.

Here is M3 schematics just for reference:
mmm_sch.png
 
jacco vermeulen

jacco vermeulen

Member
Joined 2004
gooduser said:
output stage MOSFET selection and reasons why to use IRFZ24N/IRF9Z34N pair.
Click to expand...

Not exactly (glanced through the thread rapidly)

- N and P type MOSFETs are not really complementaries
- besides input capacitance ( Cgd + Cgs for a single pair, aka Ciss), there's also transconductance.

Your IRF610/9610 suggestion in post 8, for example, has a much lower Ciss number than the IRFZ parts.
But minimum transconductance of the 610/9610 is 0.8 S.
Input capacitance figure of the IRFZ's is 3 times as much, but their transconductance is 5 times higher.

The Harris MOSFETs in Sheldon's original headphone amp are real oldies from the early/mid '80s.
(RCA/Harris' in TO-220 outfit were the first vertical MOSFETs I built output stages with back then)
They have twice the transconductance of the IRF610/9610, but 6 times the input capacitance.

If you'd like to learn a bit more of (vertical) MOSFETs (somewhat outdated, but it's free) =>
http://bitsavers.informatik.uni-stu...Books/1988_SGS_Power_MOS_Devices_Databook.pdf

(I built an SDS style headamp for my K1000)
 
Last edited:
P

pbilous

Member
Joined 2011
Paid Member
Post #22 (AMB):

As for the choice of MOSFETs, we had to consider several factors. Since this amp is going to operate in class A, we would like to choose high current devices for reliability. The Ciss should not be too high, and yet high current devices always have higher Ciss due to their larger internal junctions. Stable and assured availability is also an important factor in our consideration. Linearlity, performance and suitability for the application, and last but certainly not least, the cost is also considered because we want this amp to deliver a lot for the money.

The Hitachi 2SK213/2SJ76 series were considered. These are nice devices with low Ciss, but IMO just a bit "weak" with a Ids rating of 0.5A. The 2SK1058/2SJ162 are stronger, but the IRF devices we chose are even more robust yet with comparable linearlity, Ciss. Plus they are stocked in quantity at Digikey and Newark Electronics. The 2SK/2SJ devices are scant by comparison (at least in the US) and more expensive. The choice then became clear. The Renesas/Hitachi parts are not pin-compatible with the IRFs, so we had to choose one or the other for our PCB layout.
Click to expand...


Post #24(morsel):
Aos: AMB already addressed the Renesas MOSFET issue, but I will add that the IRF IRFZ24N/IRF9Z34N 18A MOSFETs are about $1 each from Digikey, the Renesas 2SK1058/2SJ162 7A MOSFETs are about $9 each from B&D Enterprises, and the Renesas 2SK213/2SJ76 .5A MOSFETs are about $5 each. Multiply that by 6 for the 7A devices and 12 for the .5A devices and you have $54 to $60 for Renesas MOSFETs .vs. $6 for IRF MOSFETs. MCM Electronics prices are even worse.

2SK1058 Ciss = 600pF
IRFZ24N Ciss = 340pF
2SK213 Ciss = 90pF
2SJ162 Ciss = 900pF
IRF9Z34N Ciss = 620pF
2SJ76 Ciss = 120pF

The 2SK1058/2SJ162 make no sense at all from either the cost or spec perspective. The 2SK213/2SJ76, once doubled up to compensate for their wimpy output, are too costly and space consuming despite the low Ciss. The IRFZ24N/IRF9Z34N are modern, 5th generation HEXFETs that sound good to us.
Click to expand...


Thanks for the link to the book!
 
P

pbilous

Member
Joined 2011
Paid Member
The other pair I found is the one I mentioned before, but with different P-channel MOSFET:

IRF710/IRF9610

Type
IRF710 _____ Ciss, pF = 170 ___ Id, A = 2.0 ____ gfs, S = 1.0
IRF9610 ____ Ciss, pF = 170 ___ Id, A = -1.8 ___ gfs, S = 0.9

I think this is the lowest Ciss that possible in Id>1A.
 
Last edited:
P

pbilous

Member
Joined 2011
Paid Member
And the MOSFET pair used in M3 amp:
IRFZ24N ___ C=370 _____ Id=17 ____ gfs,S=4.5
IRF9Z34N ____ C=620 ____ Id=-19 ___gfs,S=4.2

By the way, I was able to find them of eBay for pretty low price. Idk what I was looking before at, but they seem to be easily accessible to purchase.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Top Bottom