Using the THAT1646 As A Transistor Pre-Driver/Headphone Amp

Where we discuss new analog design ideas for Pro Audio and modern spins on vintage ones.
lofi
Posts: 22
Joined: Sun May 25, 2008 12:03 pm

Re: Using the THAT1646 As A Transistor Pre-Driver/Headphone Amp

Post by lofi »

Ok, so both the 1243 and the 1646 use 15V (+/-) will i get a problem if i run lots of these ( 10 amps for 5 sterio pairs) from a single source. seen crosstalk mentioned about other stuff, dont think it will apply here, but you never know ...

now to try and work out how much juice will be needed for a power supply

Iain
User avatar
mediatechnology
Posts: 5444
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Re: Using the THAT1646 As A Transistor Pre-Driver/Headphone Amp

Post by mediatechnology »

You should be able to run them on a single PSU fine.

The idle current is 160 mA/stereo pair. SInce it's heavy Class-A very little of that (maybe 2%) makes it to the load. The rest is heat about 2W. So figure 200 mA/stereo pair. Ten of them would only be about 2A max.
lofi
Posts: 22
Joined: Sun May 25, 2008 12:03 pm

Re: Using the THAT1646 As A Transistor Pre-Driver/Headphone Amp

Post by lofi »

bonza!!! jobs a goodun

Thanks Wayne
radiance
Posts: 46
Joined: Thu Jun 26, 2008 4:50 am

Re: Using the THAT1646 As A Transistor Pre-Driver/Headphone Amp

Post by radiance »

mediatechnology wrote:
Use fairly large heatsinks as the idle is about 2W/ch. With this being heavy class-A efficiency is certainly not one of its attributes.
I don't quite get this. Do you mean heatsinks for upon the 1646? Or for the transistors? Or both maybe (DOH :-))?

Thanks.
User avatar
mediatechnology
Posts: 5444
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Re: Using the THAT1646 As A Transistor Pre-Driver/Headphone Amp

Post by mediatechnology »

Hi radiance - Thank you for posting.
I don't quite get this. Do you mean heatsinks for upon the 1646? Or for the transistors? Or both maybe
Just the transistors require heatsinking and it's best all three share one. Further up the thread there's a picture:

viewtopic.php?f=6&t=23&st=0&sk=t&sd=a&start=15

The 1646 requires no heatsinking and doesn't dissipate much power. Note that the "third" power transistor, the BD139, sits on the backside. It doesn't dissipate much power either, but it's the Vbe multiplier and needs to be in close thermal contact to the output pair. Choosing a second BD139 there was a matter of convenience - it has a mounting hole and is identical to the complimentary pair both thermally and electrically. Thus, it tracks the other two, particularly the BD139 output very well.

I haven't found too much need for Vbe matching the devices either. The transistor's Vbe errors cancel to one-half their error. For example 15 mV Vbe mismatch error becomes 7.5 mV. This adds/subtracts from the 1646's common mode and differential offsets. I think the most output offset I've ever measured is about 6 mV. As JR says "YMMV."
radiance
Posts: 46
Joined: Thu Jun 26, 2008 4:50 am

Re: Using the THAT1646 As A Transistor Pre-Driver/Headphone Amp

Post by radiance »

Thanks! I'll try to incorporate it in my active mixer .......if there's enough space for the heatsinks that is.
User avatar
mediatechnology
Posts: 5444
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Re: Using the THAT1646 As A Transistor Pre-Driver/Headphone Amp

Post by mediatechnology »

You can run it at lower class-A currents if you need to but some heatsinking would still be required. Since the output runs without negative feedback it needs to be fairly high.

Someone over at Prodigy built one of these recently but I don't recall who it was off-hand. I think you'll like the way it sounds - the distortion is very, very low. Keep us posted.

Wayne
lofi
Posts: 22
Joined: Sun May 25, 2008 12:03 pm

Re: Using the THAT1646 As A Transistor Pre-Driver/Headphone Amp

Post by lofi »

bloody customs have impounded my batch, well untill i pay the £8 import duty and £16 'we payed the import duty fee for you' charge we get from the post office :x

cant wait to get to play with these things, thanks Wayne
radiance
Posts: 46
Joined: Thu Jun 26, 2008 4:50 am

Re: Using the THAT1646 As A Transistor Pre-Driver/Headphone Amp

Post by radiance »

I build the amp as part of my mixer project. It worked straight away and sounds good. I have not done extensive listening test but it seems all is ok. I'm a bit lost on which value I should set the 1K trimmers. I've set them both so I meassure 0,85V at pin 4 of the 1646. Also, I've used bd135 and bd136 because I had those laying around. I guess that's not a problem right?
Here some pics.....

Image
Image



Man, these bd135/6's get really hot after a while. This heatsink will eventually be atached to the case so I guess it will be fine....
User avatar
mediatechnology
Posts: 5444
Joined: Sat Aug 11, 2007 2:34 pm
Location: Oak Cliff, Texas
Contact:

Re: Using the THAT1646 As A Transistor Pre-Driver/Headphone Amp

Post by mediatechnology »

That's really cool to see. Nice pics!

The way I set the 1K output bias current was to measure the emitter current of either the top or bottom transistor via the I*R drop across Re. Depending on the size of your heatsink you can set the class-A current from maybe 65 mA to 85 mA IIRC. WIth 3R3 emitter resistors that becomes ~ 0.21 V to 0.28 V across the resistor. If you go to the high end of the scale the THD is very low but you might monitor the current it may take a few minutes to stabilze.

I can see from the pics that you mounted the Vbe multiplier right near the outputs so it should track really well. Being class-A it burns about 2W at idle so they do get warm.

The BD135/136 are fine. They're the 45V part and I have another application for HV parts so I just stock the BD139 BD140 and used them. For my prototype I used the hFE selected "-16" version which ranges from 100-250 at Ic= 150 mA. I don't know that it's overly critical.

I'm curious about how much output offset you see. I think it roughly works out to be the 1646's offsets which are in the 4-6 mV range plus 1/2 the Vbe mismatch of the NPN/PNP pair. I typically got 3-7 mV.

Thanks for sharing this with us.

Wayne
Post Reply