Phase splitter design

ask your general tube related questions here

Phase splitter design

Postby dhuebert » Wed Nov 10, 2004 8:39 am

Whilst reading up on long tail pair design recently, I came across some good info about the transistor version. The gain of the circuit is dependant on the collector resistor value, which is determined by DC bias considerations. To increase the gain, Rc was replaced by a current source (infinite resistance), which neatly explains all those OP amp schematics I've looked at over the years.

Question: Has anyone tried replacing Rp with a current source in their tube long tail pair design?

According to what I understand, replacing Rp with a current source makes the gain of the circuit go to infinity, this means you can go without a preamp altogether and control the gain of the whole circuit with the feedback loop... Wait a minute, a tube OP amp?

I wonder what Bob Pease would say. (see the October 18 issue of Electronic Design for an article on Philbrick K2-W tube OP amp http://www.elecdesign.com/Articles/Arti ... /8880.html )

Again I would be delighted to hear your comments.

Don
User avatar
dhuebert
KT88
 
Posts: 820
Joined: Thu May 01, 2003 9:26 am
Location: Winnipeg Manitoba Canada

Bob Pease!

Postby Shannon Parks » Wed Nov 10, 2004 6:53 pm

Bob Pease's internet show at National Semi dot com is pretty slick. They have a wicked good website.

I was just here reading up on diff amps/ long-tailed-pair here:
http://sound.westhost.com/articles.htm (look for "Current Sources...")

The noise article is what brought me there - working on a super deluxe phono stage.

OK - I've used the LM334 CCS (10mA, 40V max ratings with one set resistor - TO-92 package) in the LTP of the Eiclone. Works great. The resistance isn't infinite, but it is certainly greater than 18K. Allows me to use whatever plate resistors I'd like and get lots of drive from much lower B+ voltages. I've implemented this in a Dyna MKIII board and it works great. I have a ST70 driver board using the same circuit - just a modified Mullard ala Eico with the LM334. 340V from the ST70 will give me plenty of swing because I'm only dropping 8V instead of 140V.

Oh yeah - some claim the 'slew rate' of the LM334 can't work for audio. I certainly won't disagree with the datasheet. But the control voltage is being held constant on the LM334. I envision the CCS as an anchored pulley with the tubes doing all the work. This is my best SWAG - I certainly see no degradation in performance at full output at 20kHz up to 100kHz on my Ike with KT88's compared to a resistor LTP.

Shannon
User avatar
Shannon Parks
Site Admin
 
Posts: 3764
Joined: Tue Mar 18, 2003 5:40 pm
Location: Poulsbo, Washington

CCS

Postby dhuebert » Fri Nov 12, 2004 10:38 am

I take it from your'e reply that your'e using the LM334 in place of the tail resistor. This was an idea I toyed with when I read Aikenamps tutorial on LTP design. I thought it would be a great way to stabilize the whole circuit. I am still assembling my shop, so it will be a while before I can start prototyping, but I look forward to your comments.

some claim the 'slew rate' of the LM334 can't work for audio

I guess I'm not totally understanding. The purpose of a CCS is to be constant. If the AC impedance needs to be zero, it gets bypassed with a cap. Since the CCS is strictly a DC circuit what could slewrate possibly have to do with anything? Unless...nah, I dunno.

Don
User avatar
dhuebert
KT88
 
Posts: 820
Joined: Thu May 01, 2003 9:26 am
Location: Winnipeg Manitoba Canada

Re: CCS

Postby Shannon Parks » Fri Nov 12, 2004 3:18 pm

dhuebert wrote:I take it from your'e reply that your'e using the LM334 in place of the tail resistor. This was an idea I toyed with when I read Aikenamps tutorial on LTP design. I thought it would be a great way to stabilize the whole circuit. I am still assembling my shop, so it will be a while before I can start prototyping, but I look forward to your comments.


Exactly - as a LTP tail. It's very easy to use. As you mention, you can certainly parallel a cap with it and use it as a cathode 'resistor'.

dhuebert wrote:
some claim the 'slew rate' of the LM334 can't work for audio

I guess I'm not totally understanding. The purpose of a CCS is to be constant. If the AC impedance needs to be zero, it gets bypassed with a cap. Since the CCS is strictly a DC circuit what could slewrate possibly have to do with anything? Unless...nah, I dunno.

Don


No - you understand it perfectly. There's just a slew rate chart in the datasheet which makes the LM334 look unusable above 10kHz (ie bad for audio circuits). But I believe this slew rate refers to the control voltage changing. In a CCS configuration, the voltage drop across the set resistor remains constant as you say. The tubes are driving voltage, not the CCS so the cart is not applicable in this usage. AFAIK. ;)

There's many uses for the LM334 (like a precision thermometer circuit) so I imagine the slew rate matters in some control circuit.

Shannon
User avatar
Shannon Parks
Site Admin
 
Posts: 3764
Joined: Tue Mar 18, 2003 5:40 pm
Location: Poulsbo, Washington


Return to tube 101

Who is online

Users browsing this forum: No registered users and 10 guests

cron