advertisement


A valve line pre amp for the DIY fishy

It's a non issue IMO. The design as presented has been carefully thought out and several detail variants tried out. Most areas where one may look at it and think that something can be improved/is less than optimum are the way they are for a reason;) ie the caps are small for a reason, the PSU is not regulated for a reason etc. I'm certainly not claiming it's perfect or can't be improved but I would suggest builders stick to the design as posted as a starting point before experimenting further.
 
400v caps and RCAs are here, hopefully other bits next week. I probably start during week.
 
I've got all of the bits for a build as per the schematic apart from a 100uF instead of 47uF for C7 and just a two diode rectifier from a 250-0-250 secondary. Hopefully will get started on it this evening. Shouldn't take long.....
 
Lash-up mark 2 now working. Volts as described, 250v on top of the 30k ohm resistors and 65/68v on the ECC88 anodes.

e9dycm.jpg


Just the hum to sort out and reduce the gain a little for my applications...
 
Yes they are, I intend to mount the board vertically eventually so air flow won't be a problem. Maybe even mount the 22 ohmers there too if I cannot find a centre tapped TX..

2ioke9.jpg
de4oj9.jpg
 
Aha... Yes I was thinking about the heat rising up and cooking the caps... You've got that covered though;)
Mine was silent on the hum front but you may have to re-position the TX or experiment with wiring layout to get rid of the hum. Judged by eye I would say my TX is about another 2" further from the circuitry and it's also on the other side of the chassis.

It will be interesting to hear what you think of the sound once the teething troubles are sorted.
 
Got mine going on Sunday evening and there's no hum. I kept the transformer and rectifier about 10 inches from the circuit. All parts as per schematic apart from the split rail half wave rectification as I had a 250-0-250 transformer and a 100uF in place of the 47uF at C7. As you can see I'm not experienced with valve circuit layout! Yesterday I spent some time running individual earth wires from the star earth block to each earthed node in the circuit and even to the screen of the incoming signal wires. Even the cathode caps and resistors have separate wires to the star earth. That seemed to tighten everything up a little.





After about six hours of listening I am well impressed by the sound. Compared with my much regulated and souped up JLH headphone amp that I was quite happy with (it had ousted a B4) driving a hackernap with line transformer inputs and PMCFB1's this has a darker and more inviting/less brash sound. No listener fatigue. Better at low volumes. Vocals piano and acoustic are clearly better defined and it is more musically coherent. Doesn't have or need the Naim up front attention grabbing vitality. If I had to criticise anything it is a very very slight plumby looseness in the bass. I'm going to live with it for a while (not hard) before trying to address that. It is clearly IMHO a well tuned amp. I fear that any so called improvements may upset the balance of the sound and its coherence. A big thank you Mr Arkless.
 
No problem:) Good stuff there John and I'm glad you like it. You can "tune" it a bit by experimenting with the output cap type and value, go up to say 10uF and that would be fine, and also the two 47uF caps after the 15K resistors. Experiment between about 22 - 220uF and see what you think...

Also check that the voltages are close to as described earlier in the thread. Depending on the TX and its regulation you could be a bit either way. Within 5-10V or so of the 250V on the two 47uF caps should be fine. Increase of reduce the 15K resistors a bit to fine tune that.

Where's your vol control BTW?
 
Just measured the voltages with the primary on the 240v tap. I also have a 230v and 250v to play with.

The top of C7 is 365 volts and there's DC of 1.29v across the 100R. the two 47uF after the 15k resistors have 270v and 266v respectively. Anodes are at 79.5v and 71.5v. Perhaps things are a little high and I should switch to the 250v primary tap to get closer to the 66v on the anodes / 250v on the 47uF's after the 15k's?

I've no need for the volume control at this stage as I'm using the digital volume control on my source for now. I've got max gain by putting a 47k in place of the pot. I have taken the earths from each of the 3 filter caps straight back to the centre tap on the transformer to avoid common mode interference between the filters. I have also taken the main earth from the earthing block back to this same centre tap. Planning on a separate box for the transformer and an umbilical between.


I will try some other output caps. Currently using 2.2uF polyester. got some 4.7uF polyprops and even some 630v 8.2uF Ampohms which are way too big but will allow the removal of the zeners.

John
 
The 250V tap will get you closer so yes try that. The grounds should all go to the star block and not to the TX centre tap. The centre tap goes to C7 - and then a thick lead from C7 - to the star block.
One issue with increasing the output cap size is increased switch on thump so the Zeners should really stay. There is a switch on delay circuit for it as well but I didn't include it here as it significantly complicates things by adding as much again in SS circuitry just for the switch on delay and I didn't want to put beginners off...
 
Well done John, not so lonely now...

Fixed my hum too. Just an earth loop between my amp and the pre.
I have left just the mains earth connected to the chassis and temporarily lifted the 'star ground' off the chassis. I'll add a ground lift resistor/cap later.
No hum even with the TX closer. (Good, as mine has to be a one box unit.)

I've mitigated the power up rumble so far by switching the power amp on after the pre has warmed up, but am wondering about fitting a simple mute switch across the output for now.
 
Well done John, not so lonely now...

Fixed my hum too. Just an earth loop between my amp and the pre.
I have left just the mains earth connected to the chassis and temporarily lifted the 'star ground' off the chassis. I'll add a ground lift resistor/cap later.
No hum even with the TX closer. (Good, as mine has to be a one box unit.)

I've mitigated the power up rumble so far by switching the power amp on after the pre has warmed up, but am wondering about fitting a simple mute switch across the output for now.

Ideally the output cap would be >10uF but the issue is the switch on thump and the possibility of it becoming large enough to potentially damage the input of a SS power amp. It will always be fine with a valve power amp.

The two 47uF caps at the anodes charge up relatively slowly via the 15K resistors, and this, combined with the current limiting of the 30K anode load resistors and the Zeners, means that even with 10uF or more you should be fine... I left the output cap a relatively small 2.2uF in the original schematic just to be able to guarantee that it would be safe with all power amps. Sorry I can't answer questions on the lines of "I have an Acme MkIV power amp, will it be safe with 15uF output capacitors?"... there's just to many variables... I guess power amps with a FET input would be the most easily damaged...

In mine I have 8uF at the moment from 4 x 2uF in parallel but going up to as much as 22uF could be advantageous for sound quality, especially the bass. If keeping it polypropylene then the Zeners help in allowing lower voltage rating caps to be used. 200V rating is fine and 150 V no doubt fine but no wriggle room left (it only gets to 150V max for the time it takes the valve to warm up and then drops to 70V). 350V+ polyprops @22uF get rather large and expensive!

A simple mute switch at the output, shorting the output end of the caps to ground until it's warmed up, will do fine... providing you remember to use it!

A valve rectifier can be used and would solve the problem by warming up slowly with the ECC88.

If there is a demand for it I will publish the automatic switch on mute circuitry but it does mean considerable extra complexity, including having to rectify and smooth the heater supply so it can also power the SS delay circuitry and relay.

A pair of 12V Zeners with a pair of 1N4148's or similar could be arranged to make an output limiter but there is a chance of very slightly reduced sound quality due to modulation of the non linear reverse capacitance of the diodes. There could well be no sound quality reduction but I'm just mentioning the possibility anyway..

Another alternative, but not exactly user friendly, is to not connect it to the power amp until it has warmed up for a few minutes...
 
A bit of further info on this issue and whether to be bothered by it... Most power amps will have an input capacitor which will protect against the switch on thump. The smaller this cap the more protection is offered, so if it's 10uF (ish) or less you should be fine with no further protection to go for a bigger output cap in the pre amp.
Some power amps have input over-voltage protection but they are a minority.
A discrete BJT input stage in a typical SS power amp should sink the current involved no problem as a quick simulation shows max 3.5mA (0.8mA for 2.2uF and 2.3mA for 10uF) flowing into input transistor base emitter junction even with 22uF output cap and for about 1/2 a second only. These figures are worst case for a power amp with no input cap, no input series resistor and no over-voltage protection.

If your power amp has FET or op-amp input (ie the first thing the signal from the pre amp connects to is the gate of a FET or input of op-amp) with no input capacitor then best use switch on delay circuitry or use a valve rectifier.

To conclude on this:

1/ It's doubtful that output caps above 10uF are worthwhile anyway. I leave it to the constructor to try the subjective effects of going bigger than 2.2uF.

2/ Valve power amps will be fine whatever you do and without any protection. Obviously let the pre amp warm up before switching on the power amp in all cases, valve or SS!

3/ Most will be using a power amp with an input capacitor and probable series resistor and 80% plus will be discrete BJT at input and so you should be fine with no protection even going up to 22uF. If you are the 1 in a 1000 who's amp blows up don't come crying to me!

4/ See caveat about FET and Op-amp inputs in power amps.

5/ A circuit for a switch on delay board is available if people want it. A valve rectifier is an easy way round the issue.

6/ For experimental purposes you can fit a switch to ground the outputs for a couple of minutes while it warms up, or simply not connect the pre amp to the power amp until after leaving it a few minutes. The 100K resistors at the output are there to discharge the output caps over a minute or so if nothing else is connected to the output.

7/ There are various commercial amps around which do the same thing and don't even warn about it...
 
After about 6 hours of listening I'm rather enjoying this little preamp. I'm finding myself being drawn in to listening to whole albums on Qobuz/Tidal rather than jumping about as I normally do. The wife has also commented on how much better it sounds and even the cat was quite mesmerised by the skylarks on "Sheep" on Pink Floyd Animals Album.

I still had some volts to lose to get down to 250 volts at the top of the 30k anode resistors. I have quite a few chokes around the place and decided to fit a second layer of filtering after the rectifier and cap C7. I fitted a Hammonds 157G 30H 40mA choke two series 100ohm resistors and a 100uF capacitor to earth. The two 100R resistors were to ensure at least critical damping of this potentially resonant filter. The effect was to open out the bass and remove much of the slight vagueness I was getting in the bass. In fact it made the whole presentation more solid and precise.

I also have some Vishay MKP1848 (Metalised Polyprop) 50uF 800V caps. These are roughly 2in X 2in X 1in, so pretty large but not ridiculously so. They are really beneficial in place of C5 and C8. They have less of an impact in place of C7. My guess is that their lower ESR (approx. 5mohm) compared with that of the electrolytics (500 mohm) gives up to approx. 20db more attenuation of noise and AC rectification effects on the power rails above 30 Hz. I don't think the lower ESR will have much effect on the accuracy of the gain of the stage.

I'm always a little wary of coming to early judgement of changes such as these. Having made the changes with a theoretical justification I'm bound to be biased towards wanting to hear improvements but as I listen more I'm hearing more and enjoying more with no discernible downsides so far. In due course I'll be reversing them to see the effect.

Finally I've ordered some IRF840's and will be making a FET Cap multiplier similar to the ones on the Hackernap and the ones on the RJM audio site at this link (power supply schematic is halfway down the page):

http://phonoclone.com/diy-pho3.html

I don't really expect to hear a great deal of improvement from this but you never know. It will fit quite neatly on and is pin compatible with Hackers PFM Gyrator board.

John
 
I'd expect changes from C5 and C8 yes but not much from the rest of those modifications. Hopefully the choke will drop a few volts anyway. All good fun to experiment though:)

The output caps will have quite an effect and making these say 10uF would be something I'd be interested to get some feedback on. The 220uF caps bypassing the cathodes could be worthwhile to experiment with also and you could try values between that and say 1000uF. These only need to be 10V or more rated of course.
 
OK I'll try those next, concentrating on the effect of different values first. I have plenty of low voltage caps for the bypass and some Wima MKS4 10uF and 22uF for the output caps. A bit more tricky to change but easier than on an in situ PCB!
 
Dear Arkless,

I have tried and failed to find a modern pure mono pre-amp with only mono inputs and output. Would you be prepared to make one such based on you design here published, but using valve rectification? It will drive either a Quad II Forty or Leak 12.1 onto a single ESL. So issues of the output capacitor are nothing to consider being careful with.

Sorry for the thread diversion.

Best wishes from George
 


advertisement


Back
Top