24a Tetrode Driver THD

Having traced curves for the 24a tetrode, now is time to look at the THD performance to drive either a 45 or a 6C4C output stage as requested in the previous post.

To avoid the area of negative resistance the output signal must always be above 100V. So a quick test was made to test output performance at 100 Vpp:

  • Anode voltage (Va) = 180V
  • CCS Load (DN2540 cascoded MOSFET) sourced by 300V filtered HT line
  • Screen bias (Vg2k) = 90V from a source follower
  • Grid bias (Vg1k) = -1.7V

So anode current was set to around 5mA. Noticed that the harmonic profile does change. If current goes slightly above 5.05mA then the H2 increased and become more triode-like distortion. When reduced below 5mA, the odd harmonics increased with more pentode-like distortion.

Transition is too quick given gain of this stage and not easy to fiddle around and set the minimum distortion point.  Distortion varies from 0.60% down to a 0.50% minimum if correctly tuned:

I need to increase anode voltage to 200V and conduct further tests to find probably a better operating point. However, not bad at all for this driver so far!

 Edit (10th November 2012)

Testing a tetrode/pentode with fixed bias in my testing box is proving hard to stabilise. Where is input noise or purely the fixed bias can make the valve to drift in some minutes.

Here is my last take on Vg1k=-3V for an output voltage of 10Vrms:

The above 0.17% isn’t that great compared to some tetrode/pentodes triode-strapped as drivers such as 6P21S, 6e5p and 4P1L which distortion is below 0.05%.

Looking at the performance at full driving capability (i.e.. 100Vpp / 35.3Vrms):

You can see that performs slightly worse than biased at Vg1k=-1.7V. However the anode voltage drifted a bit from previous take. With a large gain, the fixed bias trimming of Vg1k becomes challenging in my tester.

My early thoughts are:

  • THD measurement of high-mu valves such as tetrodes/pentodes needs some sort of negative feedback to ensure the valve quiescent point doesn’t drift. Some sort of cathode bias needs to be implemented.
  • 24a is a good driver, but from a distortion perspective there are some better ones in terms of linearity.
  • THD down to 0.57% at full 100Vpp  can be achieved if bias is properly set. However in practice it is expected this to be around 0.60-0.70%
  • I would be keen to build a driver stage with a 24a and test how this sounds!

 

 

4P1L PSE tests

Testing parallel DHTs

After listening to a great incarnation of the 4P1L PSE in filament bias output stage from Andy Evans, I decided to have a look at the impact of unmatched pairs of triodes from a distortion point of view. Main reason was that when listening to Andy’s amplifier I noticed a bit of an uncomfortable treble with some strings. Perhaps the increase of odd harmonics, but wanted at least to see what was all about.

4P1L are very easy to match. you can easily get a pair with equal mu. Just randomly I picked from my collection a pair of valves with a difference of 0.5 in mu.:

4P1L PSE unmatched pair

THD is about 0.03% mainly driven by H2. It happened that one 4P1L from the pair had 0.02% where the other had nearly 0.04% distortion. The difference between H3 and H2 is about  8dB.

Then looked at a more closely matched pair (0.03 mu difference). The distortion wasn’t surprisingly different:

4P1L PSE  matched pair

Again, nearly 0.03% and difference between H2 and H3 is down to 7.5dB.

Looking at the individual performance of the 4P1L, now biased at 30mA and similar anode voltage, we can see that despite having a lower THD, the difference between harmonics is just 5dB.  This is the THD of the other  4P1L from the pair:

4P1L PSE test rig

Well, how rthis compares to a 2a3/6C4C? The latter valves are two triodes physically connected in parallel inside the same envelope. So, no matching can be done:

The previous was a low distortion 6C4C I have. Distortion is higher than 4P1L PSE, but not that much. H3 – H2 difference is about 12dB.

My early thoughts:

  • 4P1L are very easy to match
  • 4P1L PSE performs really well. Distortion of the pair is lower than a 6C4C performing at same level.
  • H3 component is higher in PSE and this could be the reason why is more noticeable when listening to strings – as I proved in practice.

 

A new 4P1L Siberian from Martin

Martin kindly sent me a couple of pictures of his latest DHT preamp incarnation of the 4P1L Siberian:

Here are some notes from Martin about his pre-amp:
“A bit more about the circuit, I have built the power supply around the nice AZ1 mesh rectifier. First cap after the rectifier is an obligato 6,8 UF oil followed by a 10H choke and a 100 UF ASC followed by the SSHV2 that I purchased via DiyAudio group buy. On your suggestion I have used the CCS from your 301A preamp. If I refer to my previous preamp with the EF6 penthode in triode connection in parafeed configuration I can say that the 4P1L preamp sounds more detailed and with more air around the instruments. Finally I will do some tests with different output caps to find out if this makes a difference. At the moment I use the Vitamin Q.

Best regards,
Martin”

Record steaming

(Tony steaming one of my records)

We had a superb time yesterday at George’s. Tony and I spent some time learning the technique from George, which proved to be very effective.

I cleaned a series of great Astor Piazzolla’s tango records that my friend Walter brought me recently from Buenos Aires. These are gems for me despite not being of a great press quality these original LPs back from the 70s are superb. Also we cleaned as well “Mingus Mingus”

My first impressions:

  1. No doubt that cleaning will remove most of noise as expected.
  2. The record clearly comes back to life. Previously it sounded dull, with no presence of  bass, piano and completely lack of brightness
  3. Dynamics are of a different dimension now.
  4. Bass and high treble are significantly recovered. You can now hear the bass and hi-hat, snares overtones perfectly well.

Simply put, this will be clearly one of my next project to be added to the list!

George’s record steaming method

Here are George’s own words:

HOW TO TRANFORM VINYL PLAYBACK FOR £20

Thee is a lot more information on a vinyl disc than we have been getting out of it. When an LP is pressed, a blob of molten vinyl is moulded between 2 metal stampers. To make sure the resulting disc does not stick when the stampers are pulled apart, the metal is automatically sprayed with mould release agent – a very fine film of grease. This dulls the sound and provides a layer in which particles become embedded, causing clicks, pops and surface noise. There are lots of machines and fluids on the market which claim to get rid of this problem, and I have tried several. None holds a candle to the method I use now – a mixture of alcohol, distilled water and steam.

After seeing some videos on YouTube and reading some interesting forum threads, I bought a hand held steam cleaner on Ebay for £30. I got a Vax S4 because they also did well on upholstery in the Which? Tests, I see that the Robert Dyas shops are selling a basic steam cleaner for £20 which will do the job required. I use purified water in it for cleaning records.

METHOD

I have a Moth Record Cleaner machine built from a kit years ago. You clamp a record to a 4 inch turntable driven at about 15 rpm by a high torque motor. You apply cleaning fluid with a brush, flip the record over and vacuum it from below using a slotted tube covered with pads of velvety felts (see http://www.britishaudio.co.uk/mothrcm.htm) .

I have added a 12 inch platter and used some B&Q plastic plumbing tube and bends to bring the vacuum tube up to a level where I can vacuum a record from the top. This means I can do several different cleaning runs without having to unclamp and flip it several times. My platter is a circular hardwood varnished chopping board, with a threaded insert in the spindle hole and a slight depression sanded down for the label area. I protect the label with a plastic disc (easily made from a lid. The threaded insert allows me to bolt the record down firmly.

I first spray on (using an atomiser from Boots) a mixture of 30% isopropyl alcohol, 70% distilled water with a drop of photographic wetting agent. This gets worked thoroughly into the grooves of the rotating record, working the brush both against and with the direction of rotation. The amount of liquid used depends on how wet the brush is – the first one needs a double dose to “load” the brush – and how dirty the record is. I do not vacuum it off at this point.

Next I spray the steam at an angle across the rotating record from about 2 inches, moving the nozzle all the time, letting the platter go around twice. |The melting point of vinyl is well above the temperature of the steam and you won’t damage the vinyl provided you keep both vinyl and nozzle moving. I have no qualms about using this on some very valuable collectors’ items.

Then I brush again as before, both with and against the grooves. (At the moment I am waiting for more brushes so I can use a different brush for each stage.) The liquid is then vacuumed off. I then spray it again with steam to rinse it and then vacuum it thoroughly. The velvet material on the vacuum tube is cleaned with a quick blast of steam every few records. I unclamp the record and blot up any water around the label with tissue and afrer doing the other side leave it for few minutes to dry off.

RESULT

Tony Rees came round with some LPs and after I cleaned one said that it was like seeing a painting which has had its old varnish removed so that for the first time you really saw what was there. That about sums it up. I was staggered at the amount of subtle detail I had been missing. Records I have owned for 30+ years including one cleaned on a top-of-the-range Keith Monks machine were transformed. Charity shop mistakes became listenable. Scratches were still there but surface noise was much less. It was like a major cartridge upgrade – and it cleans stains off upholstery.

CONCLUSION

Mould release agent is very persistent and dulls the playing potential of the LP.

The combination of cleaner and steam loosens it and its embedded particles and it can be cleaned off, revealing the vinyl for the first time. I have cleaned old, new and neglected LPs and some previously cleaned in other ways. All were improved, some beyond initial expectations.

Anyone wanting to hear the result on a few of their LPs is welcome to contact me.

Cheers

George”

THD, Gm and μ tester

As part of my repairing of the curve tester, I had to do some changes to the transconductance (Gm) meter section. Currently I’m leveraging most of the curve tester to also measure Gm, μ and distortion (THD). Albeit the latter is rarely used as I prefer an external equivalent CCS circuit that is not inside the tester as the output signal comes out cleaner. The curve tester provides all sockets, HT power adaptor, meters and bias supply.

This is my latest circuit:

The additional protection to fuse (F1) is the diode D2 which can protect the LCD panel meter A1 in case of an unexpected anode short. P1 and R4 were chosen to allow a precise setting of the anode current at low levels and some protection to the CCS when P1 is set to zero. M1 is bolted to chassis and is carrying all the effort when providing current at lower anode voltages. M2 on the other extent can be a TO-92 type. R1 was added to allow a bleeding path to C2 when not measuring transconductance. The bias section is a simple adaptation of Merlin Blencowe’s “Power Supplies for Tube Amplifiers”, which I suggest you take a look at as Merlin covers very well the most common valve bias circuits

With this circuit I can measure very accurately transconductance at any desired point.  I highly recommend you Alan Douglas’ “Tube Testers and Classic Electronic Test Gear”, which has a lot of details around how classic valve testers work, challenges around Gm measurement and obviously some good ideas and suggestions for calibrating and measuring Gm correctly.

Obviously adding an amplifier to the Gm tester section could improve the accuracy of low transconductance valves. But that would be for another time!

Listening to the 26 preamp

Now that I have a very quiet preamp indeed, I’m a happy bunny. Changed today the pair of White Westinghouse 26 ST NOS that I’ve been using for over a month with a pair of RCA CX-226 and have to say that I enjoyed a far more sweeter and rounded tone out of this preamp.

Bass is still deep and treble is clear with the sweetness of the globe valve fingerprinted in the tone as you would expect. I will run this setup for a month and report results…

You can still see the clips used to provide earth return at the output to avoid the ground loop that I unnecessary created. God knows how may tweaks I did to improve the HT rail ripple before I realised that it was a simple ground loop.

Glow tubes fed by a CCS are not just very quiet, they are really cool

955 / VT-121

 

After a long delay, I finally did what a friend asked me. Test the lovely 955/VT-121. This acorn-type RF valve has a 6.3V filament, 1.6W maximum anode dissipation and 250V maximum anode voltage.

As per Paul’s recommendations, I looked at minimising distortion around 6mA/250V.

Here is the distortion using a DN2540 Cascoded CCS load driving the output to 10Vrms:

And here is the distortion at 1 Vrms output level:

This little fellow is a very linear valve at low signal levels and still performs really well at 10Vrms output when THD is only 0.11%. A great candidate for a preamp so would be keen to read some comments around the sound of it…

Edit

Again, as per Paul’s request, I looked at  what the THD was for lower biasing points. The operating point was set for Ia=6mA and distortion measured for small signal, i.e. output voltage of 1Vrms as shown in the table on the left.

It’s very interesting to see that there is a minimum at -1.4V Below this grid current effect increases the distortion sharply as expected:

 

What is more interesting though, is the harmonic distribution change. H2 is reduced whilst H3 it is not. H2 and H3 levels are equal. Wonder if this is the residual harmonics from my oscillator 🙂