It is currently Mon Nov 20, 2017 8:27 am

Homebrew amp

Domestic Valve & Transistor receivers, Radiograms, Gramophones, Amplifiers, Hi-fi, Speakers, Record players, Music centres, Tape machines, Cassette players & Jukeboxes
 
Posts: 761
Joined: Sat Jan 07, 2012 10:21 am
Location: Sleaford lincs

Homebrew amp

Post by Dr Wobble » Thu Dec 31, 2015 8:09 am

I've thrown myself in the deep end with this amplifier and going about it all wrong, but nothing ventured. I've been meaning to build a valve amp using 807's for a while mainly because they look nice and are cheapish in comparison to other OP valves. That and Miguel was building one so we said we'd build them in parallel.

Looking for a suitable OPT, I bought a Hammond 1250TA off Ebay cheap but it has a low primary impedance and is capable of 120w. This makes my third valve amp build not the easiest. I'm mainly building it as a learning experience, am not bothered about low distortion, I want a bit of distortion, what's the point of using valves if the output isnt coloured.

Here's where I've got so far - see schematic ( theres a few mistakes in the drawing ) and pics. I have a problem with the second stage which is distorting but have an output albeit a puny one. I'm beefing up the power supply today, I'll use my big iso tfmr and variac.

Ta for looking, Andy.
Attachments
schematic amp.jpg
first schematic- base for building on
Amp 3 003.jpg
Amp 4 001.jpg
amp 7 005.jpg
wires everywhere
amp 7 001.jpg
output top input bottom

 
Posts: 719
Joined: Tue Jan 27, 2015 2:02 pm
Location: Chatteris, Cambs

Re: Homebrew amp

Post by PYE625 » Thu Dec 31, 2015 9:44 am

I'm impressed....great fun to be had by building your own :aad

 
Posts: 472
Joined: Wed Aug 17, 2011 5:03 pm
Location: Aylesbury, Bucks.

Re: Homebrew amp

Post by GlowingAnode » Thu Dec 31, 2015 11:11 am

Hi Andy, glad to see you've made progress since I saw you.
Whilst you are developing your amp, I'd recommend running the output valves (807's) in auto-bias rather than fixed bias.
Insert a resistor of value 47 ohms between the common star point of the 6 off 10 ohm cathode resistors and chassis 0v.
This resistor will dissipate a few watts (6 W) so needs to be adequately rated, note it should not be decoupled with a capacitor (as per Williamson design).
The grid leaks now connect to chassis 0v rather than the slider of the bias pot.
With Ik of 60 mA per valve, the cathode will be at +17v (ish) wrt g1.
Leave the 10 ohm resistors connected, as these provide a handy means of checking Ik.
As previously discussed, 4 off 807's would be just about adequate, so running 6 off will make the design easier and driver stages less critical, and should loaf along nicely.

Cheers, Rob.

 
Posts: 2714
Joined: Sun Sep 01, 2013 6:19 pm
Location: Behind the sofa

Re: Homebrew amp

Post by Cathovisor » Thu Dec 31, 2015 1:16 pm

Dr Wobble wrote:Here's where I've got so far - see schematic ( theres a few mistakes in the drawing ) and pics.

I can see one - how you're feeding the screen of the EF37A.

However, it looks good so far and as for the distortion - well, you're talking to a heretic here who thinks the phrase 'valve hi-fi' is an oxymoron anyway! :qq1

Also, good advice from Rob. Try it out, you have nothing to lose!

 
Posts: 4203
Joined: Wed May 02, 2012 3:35 pm
Location: Worksop

Re: Homebrew amp

Post by Refugee » Thu Dec 31, 2015 2:58 pm

It looks to me like the screen wire to the EF37A has a short to the top of the volume pot.
If this error has been carried through to the physical circuit the second stage would be driven hard on.

 
Posts: 472
Joined: Wed Aug 17, 2011 5:03 pm
Location: Aylesbury, Bucks.

Re: Homebrew amp

Post by GlowingAnode » Thu Dec 31, 2015 5:42 pm

Forget the first two stages for the moment, feed the input directly to the phase splitter via 0.005uF ( sounds a bit small, try 0.05uF) coupling cap.
Once this stage is working progress back towards the input.
BTW is the 250k connected to the grid of the 6SN7 a volume control? If so it's no good putting it inside the nfb loop.
Rob.

 
Posts: 224
Joined: Fri Jul 13, 2012 4:39 pm
Location: Dorset

Re: Homebrew amp

Post by Marconi_MPT4 » Thu Dec 31, 2015 6:32 pm

Hi Andy,
This amp should perform very well and 807 is a good choice.

As Rob mentioned in previous post, the volume control has to be located outside the feedback loop. Adjusting volume changes the forward path gain , while the negative feedback 'gain' is fixed. This will upset relationship between the two and may result in instability or indifferent frequency response variation.

Phase splitter anode load resistor values 120k/82k are very different giving unequal 807 drive.

Firstly I would suggest temporarily disconnecting the NFB connection to check out open loop performance because feedback can sometimes mask problems within individual gain blocks in the forward path. Three capacitively coupled stages might, depending on their time constants, result in motor-boating due to becoming a phase-shift oscillator.

Great project.

Cheers
Rich

 
Posts: 472
Joined: Wed Aug 17, 2011 5:03 pm
Location: Aylesbury, Bucks.

Re: Homebrew amp

Post by GlowingAnode » Thu Dec 31, 2015 7:14 pm

My suggestion of feeding the phase splitter directly was based on the assumption that the nfb has been omitted at this moment in time.
Rob.

 
Posts: 761
Joined: Sat Jan 07, 2012 10:21 am
Location: Sleaford lincs

Re: Homebrew amp

Post by Dr Wobble » Fri Jan 01, 2016 8:24 am

Thanks Rob, thanks all for your comments and advice. I'm going to remove the second stage as it's doing nothing ( If I turn the volume up above "1", the signal is clipped with flat tops, increasing HT doesn't fix this and the bias point is about right.) The first two stages I copied from an 807 data sheet, just to get something to play with. The mistakes in the the drawing are 1) cathode/g3 connected to anode EF37A, 2) g2 EF37A connected to pot. 3) dotted pot should be to the right and not in cathode bias circuit. When I get the amp sorted out, it represents a pot (2k) to feed NFB; there is no NFB connected at present. G2 cap will then be connected to cathode not earth. The 250k pot isn't in the proposed NFB loop, its a volume pot between the EF37A anode and g1 of the following 6SN7.

I take your point about autobias Rob, I had -30v yesterday with .6v on the cathodes. The 20k pots don't have enough adjustment so at some point will change to 100k.

Mike, completely agree. I'm making this amp mainly as a learning experience and I actually want a bit distortion, hopefully mainly 2nd harmonic.

I have a problem with the OP 807 stage loading down the LTP. The LTP has 40v RMS on each anode but is about 1v @ g1 on the 807's. I'm thinking of putting a AC coupled cathode follower in-between to stop this. The anode R's were both 82k on the LTP to start with, but that resulted in a disparity in the out of phase signal; it was quite a bit smaller. I used a pot to balance them, read off the value and replaced pot with 120k R. Both signals are now within about 3v of each other. I can fiddle with this later on.

Too many phases is something I'm worried about Rich. I'm winging it as it is. I can draw a load line, calculate gain etc but phase compensation, LF/HF filters and all that jazz are beyond me at this point.

I've been reading a lot and have piles of printed off books upstairs and have learnt a lot in the last few weeks. It helps having a real application to relate to, when learning theory. One gripe I have with all the books or online info is that they treat each subject in isolation. EG here's a common cathode/cascode/PP stage but don't mention their interaction with each other. This is also an omission in solid state amp literature . Another problem I've encountered is authors saying A, B or C then suddenly jumping to F. I threw down (gently) The Art of Electronics the other night when trying to learn about Thenevins Theory and the buggers don't give you the answers to the questions at the ends of the chapters. I'm going to write to Mr Hill and Mr Wotsisname about this! : )

Thanks for your interest, Andy.

 
Posts: 224
Joined: Fri Jul 13, 2012 4:39 pm
Location: Dorset

Re: Homebrew amp

Post by Marconi_MPT4 » Sat Jan 02, 2016 3:54 pm

Hi Andy,

I read many years ago that The Art Of Electronics by Horowitz and Hill was essentially written for post graduate engineers to bridge the gap between practical application and theory of electronic principles. They did produce a separate volume with answers to questions which may be available on 'tinternet. Slightly off topic... Recently when raw graduates were interviewed for a hardware design position, I was shocked at the lack of practical ability demonstrated by degree students when asked to complete a simple test sheet. First question was a very simple five resistor network requiring nothing more than an understanding of Kirchoff or superposition theory. Only one out of twenty applicants managed to partially complete the question. HNC/HND students fared much better. Back on topic.
Dr Wobble wrote:I have a problem with the OP 807 stage loading down the LTP. The LTP has 40v RMS on each anode but is about 1v @ g1 on the 807's. I'm thinking of putting a AC coupled cathode follower in-between to stop this.

If the bias on each 807 is -30V you can not apply more than 2 x 30V Peak or 60V peak to peak and stay within class A operation. Now, if correct, 40V RMS is roughly 110V peak to peak and is way too high for class A. With a signal this large either the grids are being driven negative into cut off, but not above 0V (AB1) or additionally the grid become positive and start drawing grid current (AB2). This might account for signal degradation.

An 807 can handle these modes of operation, but the driver has to have a low output impedance for AB2 and STC data states an 807 requires a total grid circuit impedance of less than 500 ohms. Grid G1 becomes an anode when driven positive and current flows hence the need for a low impedance driver. Traditionally transformer coupling provides the correct drive impedance for this operation. Normally unless the design is operating in AB2 you do not need cathode follower drivers. What is the design criteria for your output stage and chosen class of operation? It is essential to set grid bias and drive voltage for any given mode as per valve manufacturers data.

What test frequency is being used, guess its 1kHz? With a grid leak resistor of 100k it is usual to expect a coupling capacitor value of between 0.22uF and 0.47uF. However this would not account for 40V RMS at the anode being reduced to 1V, something else is going on here. In class A operation the output valves should offer very low loading for the driver output signals. Are the 807 valves OK and reasonably matched or just simply being seriously over-driven?

One way to reduce the coupling capacitor count is to remove 6SN7 stage completely and DC couple the LTP input to the EF37A anode. You will have to adjust DC conditions for both stages but effectively only one capacitor to consider in feedback loop phase margin calculations.

Cheers
Rich

 
Posts: 2714
Joined: Sun Sep 01, 2013 6:19 pm
Location: Behind the sofa

Re: Homebrew amp

Post by Cathovisor » Sat Jan 02, 2016 4:07 pm

Here's an off-the-wall suggestion: if the 807s need a low-impedance drive to their grids, how about using the 6SN7 as a dual cathode-follower fed from the anodes of the 6SL7 acting as the LTP?

Just a thought...

 
Posts: 390
Joined: Sat Jun 15, 2013 10:59 am
Location: Near Glasgow

Re: Homebrew amp

Post by Niall » Sat Jan 02, 2016 4:08 pm

Marconi_MPT4 wrote: HNC/HND students fared much better.


I have to ask, did any of them get the job? Because I have been in the same situation from the other side, and in the end it was "sorry, other candidates had degrees..."

(So I went to get one of those, and the University managed to effectively destroy any interest I had in the subject for many years.)

 
Posts: 472
Joined: Wed Aug 17, 2011 5:03 pm
Location: Aylesbury, Bucks.

Re: Homebrew amp

Post by GlowingAnode » Sat Jan 02, 2016 8:31 pm

Further Rich's comments, is there a 100k on each g1? Ie is the phase splitter trying to drive 3 lots of 100k in parallel (33k)?
Rob.

 
Posts: 761
Joined: Sat Jan 07, 2012 10:21 am
Location: Sleaford lincs

Re: Homebrew amp

Post by Dr Wobble » Sun Jan 03, 2016 8:55 am

Yes Rob, the resistance on both is g1 to ground 38k g1 - g1 67k. I understood this to be low impedance and thought this is what is pulling the signal down.

Thanks Rich, that helps explains what is going on in the OP stage. I havn't quite developed this skill yet and have had trouble with the data sheets as they don't give a sheet for the conditions I intend to run the amp in, IE class AB1 UL 300v HT (maybe 350v). If I'd chosen the 6L6 or EL34 as my OP valve, I would have a lot better datasheet to go on. There is a data sheet for 300v on the screens, but all the detailed bumf for the 807 has the anode at a higher voltage the the screen, something like 500v anode, 300v screen. I can't do this as I'm running in UL mode.

I've tried to build a cathode follower first using a 6V6 triode strapped. I chose this rather than a triode because in theory a pentode should make a good CF as it has high gm/current gain. Whilst this worked to some degree, I had a big problem trying to get the bias right. I started by using an R above the cathode R to get -10v on g1, but ended up with -30v which obviously is no good. Anyway, and lots of experimentation, I got nowhere; the signal was distorted and had a square wave on top of it, so something very wrong. Oh and the PS stopped working.

So I'm going to go back to the drawing board and re do the PS in light of this and what Rich has said.

Thanks all, Andy.

 
Posts: 2516
Joined: Sun Jun 10, 2012 7:23 pm
Location: Basildon

Re: Homebrew amp

Post by crackle » Sun Jan 03, 2016 9:34 am

Would the 0.1uf cap on the grid of the second triode in the phase splitter give an AC short to ground?
And what does the 500k and .1uf do on the anode circuit of the EF37A, is this to stop HF instability?
Mike

 
Posts: 761
Joined: Sat Jan 07, 2012 10:21 am
Location: Sleaford lincs

Re: Homebrew amp

Post by Dr Wobble » Sun Jan 03, 2016 11:30 am

Mike the 500k on the EF is to drop g2 down a bit, sorry, not clear in drawing.The cap is to short AC to ground, I think to stop instability.

I've found why the phase splitter isn't working, some doofus wired it up wrong with the cap on the wrong side, it is as you say to short AC to ground. I've looked again at my figures and measurements for the PS and somethings wrong, the bias is too low. I've thought about direct coupling the EF to the PS, but can't get my head round at present. The DC on the anode of the EF is 109v, the bias on the PS should be at -2 so therefore if the cathode is at 111v, that would be right. But as Rich says the PS has too big a output, so I need to sort that out too.

I'm wandering whether to scrap this version and use the pre stage from this instead - see schematic. Before I do, I'm going to try and get the PS working and do some sums for the KT88 amp.
Attachments
kt88_6a.gif

 
Posts: 761
Joined: Sat Jan 07, 2012 10:21 am
Location: Sleaford lincs

Re: Homebrew amp

Post by Dr Wobble » Sun Jan 03, 2016 12:33 pm

Just done the load line for the 6SN7 PS in the KT88 amp. Two things, the 6SN7 datasheet says that max anode voltage is 300v, the HT is 450v, though if I've calculated right the quiesant I with mid way bias @ -10/12v will give anode V/I of 250v/1mA. One question which gives more output swing, a low or steep loadline?

Andy

 
Posts: 761
Joined: Sat Jan 07, 2012 10:21 am
Location: Sleaford lincs

Re: Homebrew amp

Post by Dr Wobble » Mon Jan 04, 2016 8:33 am

Did a bit more work on the amp yesterday. I got the PS back up and running and did as suggested by Rob, put the test signal straight into it. I noticed that the two signals went out of balance as the amplitude of the test sig went up. I changed the 120k anode R back to a 82k and all seems well.

I changed the bias on the OP section to auto bias by using a 7w 300 ohm as a centre "post", and then connected the 20k trimmers direct to the cathode with the six 10r sense resistors connected to these and in turn these were connected to the 300r center post, giving 330 ohms on each cathode. Hope that makes sense. I adjusted each 807 cathode V to 60mA. Oh, I disconnected four of the 100k g2 resistors too. It got too dark to see after that so couldn't take any readings

Connected it all together and got a nice but low sine out. I'm going to push it a bit more today and see what it will put out and take some readings.

Andy.

 
Posts: 472
Joined: Wed Aug 17, 2011 5:03 pm
Location: Aylesbury, Bucks.

Re: Homebrew amp

Post by GlowingAnode » Mon Jan 04, 2016 9:30 am

Dr Wobble wrote:
I changed the bias on the OP section to auto bias by using a 7w 300 ohm as a centre "post", and then connected the 20k trimmers direct to the cathode with the six 10r sense resistors connected to these and in turn these were connected to the 300r center post, giving 330 ohms on each cathode. Hope that makes sense. I adjusted each 807 cathode V to 60mA. Oh, I disconnected four of the 100k g2 resistors too. It got too dark to see after that so couldn't take any readings

Andy.


Morning Andy, glad the ps is starting to behave itself.
300 ohm sounds way too high for a common cathode resistor for 6 valves, I estimate it should be nearer 47 ohm.
How and why are you adjusting the bias, the whole point is that it should autobias without any need for adjustment? I don't really follow, a diagram would really help please.

Cheers, Rob.

 
Posts: 761
Joined: Sat Jan 07, 2012 10:21 am
Location: Sleaford lincs

Re: Homebrew amp

Post by Dr Wobble » Tue Jan 05, 2016 8:14 am

I've gone back to the drawing board and scrapped the input and phase splitter because the gain was too high (an EF37A is a bad choice here), to be honest, I'm not sure I even need an input stage for line level sources because I only need 45v P-P for g1 - g1, so a LTP is quite capable of this on it's own. I did some sums last night to try and get my head round the OP stage.

Looking at my load line and the data sheet and having listened to what Rob and Rich have said we have 300v HT, with a quiesant current of 60mA @ -22v bias, thats in the class A part of the load line so to auto bias that would need a cathode R of 360r for one valve or 330 is nearest, am I right Rob? * This should give me approx
14w per two valves. Using 350v will give a bit more power, but is over the 300v screen rating but I've read the 807 should handle it.

* What I didn't take into account before was that if an R is shared between 2 valves, current doubles, value halves and power rating goes up. this setup is complicated by having 6 valves, so if we have one cathode R we would need a 55r or 47r, which is what you said Rob. However, I havn't a big 47r resistor so used a big 300r shared between all 6 as that was all I had and I forgot about parallel resistances, so thought my setup was right, which it isnt. I have 60m. My thinking behind using trimmers was so I could balance the valves up as theyre not matched. See rough schematic.

I've used a 6SN7 common cathode triode stage as input valve, HT 240v, 33k anode resistor, 560r cathode R with 50u bypass cap biased at -3v with an input of 1khz .5v P-P gives me about 20v ouput which goes through a 1 meg pot direct into g1 of a 6SN7 LTP PS. I'm using 47k anode R's biased at -7v with a 1meg grid leak R connecting g1 to g2, g2 is AC coupled to ground by a .1u cap, HT at 300v. This PS isn't working however, because I think the cathode R ( CCS) isn't big enough at present, I need to do some adjustment.

I'll have more to report later when I've altered the circuit. Andy.
Attachments
Cathode resistor bog up.jpg

 
Posts: 472
Joined: Wed Aug 17, 2011 5:03 pm
Location: Aylesbury, Bucks.

Re: Homebrew amp

Post by GlowingAnode » Tue Jan 05, 2016 6:29 pm

Andy, remove the 20k pots, they are not necessary.
Use a common cathode resistor as close to 47 ohms as you can find. If necessary parallel up to achieve this value. This will dissipate around 8 watts so needs to be adequately rated.
As you correctly say the dissipation will increase with the number of valves.
Power up and check the current sharing between valves by measuring the voltage across the 10 ohm resistor.
As I explained to you previously you need to get the static conditions right first.
Cheers, Rob.

 
Posts: 761
Joined: Sat Jan 07, 2012 10:21 am
Location: Sleaford lincs

Re: Homebrew amp

Post by Dr Wobble » Wed Jan 06, 2016 3:48 pm

Just finished testing the amp and for a first go, it's not too bad. Before onset of clipping with a dummy load of 8 ohms I got 40v ish P-P or 16v RMS with a 1khz sine. I tried it with a speaker/resistor combo but it was too loud. There was hum, but that's to be expected with the "layout" as it is. I'm quite chuffed to be honest, my first homebrew amplifier designed and built by yours truly with a little help from my friends.

Of course this bit is the easy ish bit, to fine tune it and apply NFB and test for distortion and wot not and then build it into a proper amp will be less of a doddle.

Andy.

 
Posts: 472
Joined: Wed Aug 17, 2011 5:03 pm
Location: Aylesbury, Bucks.

Re: Homebrew amp

Post by GlowingAnode » Wed Jan 06, 2016 6:27 pm

Well done Andy for sticking with it.
30 watts is not bad, what anode / screen voltages are you running?

Cheers Rob.

 
Posts: 761
Joined: Sat Jan 07, 2012 10:21 am
Location: Sleaford lincs

Re: Homebrew amp

Post by Dr Wobble » Thu Jan 07, 2016 7:25 am

Thanks Rob, it's 32w "RMS" or 200w P-P which brings us to the old " How powerful is my amplifier?" chestnut.

It's running at 60mA quiesent 310v on anodes and screens at - 20v bias. Datasheet says 88mA no sig, 138mA peak I think, so it should kick out a bit more.

I'll post a schematic later and do some more testing later. Andy.

 
Posts: 761
Joined: Sat Jan 07, 2012 10:21 am
Location: Sleaford lincs

Re: Homebrew amp

Post by Dr Wobble » Fri Jan 08, 2016 7:12 am

Schematic as promised. I really must take time to suss out a PC schematic program, but it's easier and quicker to draw it. At some point I'm going to see how the PS behaves with a CCS, hence the transistor. This circuit will need some work.

Mike came over yesterday and cast his expert eye over the amp, no howls of laughter, thank bog. We were trying to find the cause of the hum which is quit significant at full whack.

Andy.
Attachments
CCF07012016_00001.jpg
Version two, homebrew 807 amp 16

Next

Return to Radio & Audio



Who is online

Users browsing this forum: No registered users and 3 guests