SRPP stage useing a 5670

[Nick]

ah or do you think it draws more than 6mA under load - how is that CLASS A then

Ok, but my definition of class A is that current is constantly drawn, not that a constant current is drawn. But I agree it does have to come from somewhere. I am just arguing a silly point.

[JamesD]

Does this make sense? With 6 mA quiescent current it could in theory in Class A swing from 0mA to 12mA so its not constant current but in practise the active device needs some headroom to stay in its linear region... so higher distortion as it gets nearer the current swing extremes and this swops between bottom and top device with the polarity of the voltage swing but as this is PP Class A top and bottom device are both conducting and the output is the sum of the current from both devices so when, say, the top device is conducting towards 12mA it is as linear as it gets but the bottom device is near to 0mA and is as non-linear as it gets - so in the load is the sum of a large signal low distortion signal with a small signal high distortion signal - the question then is if this leads to higher distortion overall? Experiment suggests it does as SRPP is known for poor distortion performance at high signal levels... to what degree is this true for SRPP powered from a single rail?

[Nick]

Yes, but I could suggest a more complex setup with helper split rails that draw no current at idle but does get involved on signal excursions. Sort of like current dumping. Again, not of any real practical use, but some idle thinking.

[ruffrecords]

An SRPP is a true push pull stage so it can swing plus or minus twice the quiescent current. With 6mA quiescent it can swing 12mA peak or more than 8mA rms which is 20V rms into 2500 ohms.

Cheers

Ian

[ruffrecords]

OK, now I am back at a proper (Linux) computer without hard wired control key sequences, here is the long version.

When I started out in electronics in the late 50s everything was valves. I built the usual power amp, a tape interface for a Brenell tape deck and various radios. By the time I left school and went to university at the end of the 60s transistors were everywhere We were the first university intake that wasn't taught valves. I had always had an interest in audio and recorded my friends from school on the Brenell and at uni I recorded lots of bands at the Open University studios for inclusion in the university weekly program on Nottingham local radio. In the mid 70s I worked for a few years for Neve designing audio mixers but the pay was not so good ( and I had a family to support) so for the rest of my career I designed microcontroller systems. But I kept my hand in recording demos for local bands using various reel to reel machines and mixers I mostly built myself..

I was lucky enough to make some money from mobile phones and 20 years ago I retired at the ripe old age of 50 .Since then I have returned to my first loves, valves and audio mixers. I soon discovered that the mixers I had designed at Neve back in the 70s were now considered vintage and , along with lots of tube gear from preceding decades, were generally revered for their vintage sound. So I decided to start by designing a valve microphone preamp. I researched loads of classic designs and also read lots about what people expect from valve designs these days. I soon discovered that negative feedback (NFB) was frowned upon by many so I looked for a topology that could provide a reasonable gain with low distortion and no NFB. That was when I discovered the mu follower. I then searched for tubes with low intrinsic distortion so that even without NFB the distortion would be acceptable. Up popped the 6SN7 and its more conveniently packaged B9A version the 6CG7. My first mic pre, and six channel mixer was based on two stages of 6CG7 mu followers. The mixer is still in use in a Swiss studio. The mic pre was lent to a guy in the USA who managed to allow it to be stolen by a rock band

The big problem with mu followers is their drive capability. In more complex mixers you need to be able to drive several buses and a direct output and to provide adequate headroom it needed to be capable of at least 10V rms output into a load as low as 2400 ohms. The 6CG7 mu follower cannot do this. So I needed to find a toplogy, and maybe a different tube, with greater drive capability. As part of the first mixer design I needed to provide a headphones amplifier. I discovered the Pete Millett design using an ECC99 (I think) based SRPP and realised this topology had the drive capability I needed. But the raw SRPP produces several percent distortion far to much for professional use. For some reason Pete Millett did use any NFB but I felt I had no choice. After some experimentation I settled on a design that closed the NFB loop right down to dc so there were no LF stability problems and the distortion was reduced to a tenth of its original levels.

The other problem with the 6CG7 and ECC99 is they are so tall. I wanted to squeeze more channels into my mixer designs so I needed a shorter tubes, the same size as the 12AX7. Enter the ECC88. Higher mu than the 6CG7, higher gm too and much lower rp which meant it was quite capable of driving a 2400 ohm load to the levels required when configured as an SRPP. Distortion was still an issue but I knew how to cure that from the work done on the headphones amp. I also managed to tweak it further so that the closed loop gain could be varied over a wide range simply by changing the value of one resistor. Half a 12AX7 and one 6922 make a complete amplifier capable of up to 40dB gain. Two of these (three valves) on one PCB form my now standard twin llne amp.

Long story short, if you want to design and build valve audio mixers (and any other type of valve based pro audio gear) , the SRPP is an indispensable tool.

Edit: For more info check out the DIY tab of my web site https://www.customtubeconsoles.com/ where you will find lots more info.

Cheers

Ian