Zog Audio: Malakron's JX-8 Hyperdrive

What people are working on at the moment
Post Reply
AndrewR
User
Posts: 44
Joined: Tue Oct 02, 2007 2:00 pm

#1 Zog Audio: Malakron's JX-8 Hyperdrive

Post by AndrewR »

Well that's what I'm calling it :p

My latest project. A unique valve amplifier design all of my own - with a nod towards the interstage push-pull 6B4G design on the Meta-Gizmo site. But it's rather more than just that...

It's two stages of an ECC99 directly coupled together, the last of which has an interstage transformer on the anode that splits the signal in two for the pair of 6B4G grids.

What does that mean? Well it does not use any capacitors in the signal path for starters. Also the interstage transformer neatly divides the signal to the class A push-pull stage.

The power supply is also most interesting, and designed in a way that increases the agility of amplifiers handling of the music over the spectrum. Here I'm calling it Hyperdrive technology - something that would give marketing guys wet dreams. Basically it uses descending values of supply caps, which works a treat! Sadly I think the use of motor run capacitors (from Ducati!) was a bit of a wrong-turn, and are probably difficult to take out.

However I am loving the BHC capacitors - IMHO a far better and more practical choice than Black Gates (from previous project). Works well with the tantalum resistors too :) The iron is from Sowter.

The end result is superb! Very open and rythmic with plenty of bottom end welly. It has the capability of grabbing attention (in a positive way) even with 8W running into 85dB/W/m Katans.

Andrew Randle

P.S. I wouldn't recommend using the name "Hyperdrive" in a product context, as it's already trademarked by Chang Lightspeed power conditioners :(
Attachments
ZOG2.jpg
ZOG2.jpg (147.39 KiB) Viewed 11084 times
ZOG1.JPG
ZOG1.JPG (132.29 KiB) Viewed 11084 times
User avatar
Mo
Old Hand
Posts: 256
Joined: Tue Aug 07, 2007 10:24 pm
Location: Somewhere

#2

Post by Mo »

Hey Andy, welcome to the forum. I'm sure you'll be welcomed by others too :)
User avatar
pre65
Amstrad Tower of Power
Posts: 21373
Joined: Wed Aug 22, 2007 11:13 pm
Location: North Essex/Suffolk border.

#3

Post by pre65 »

Hi-a nice piece of kit! I thought of doing my PT15 monoblocks "mirror image" but it was too much work (for me).

Motor run caps are good (and inexpensive) !


ps Why are they good ?
The only thing necessary for the triumph of evil is for good men to do nothing.

Edmund Burke

G-Popz THE easy listening connoisseur. (Philip)
User avatar
Nick
Site Admin
Posts: 15711
Joined: Sun May 06, 2007 10:20 am
Location: West Yorkshire

#4

Post by Nick »

Hi, nice looking amps,
Basically it uses descending values of supply caps, which works a treat!
Interesting, have you any more details on that, power supplys are one of my areas of interest, Andrew and myself have had some interesting (sometimes beer fulled) discussions on them (including the prime number theory, but thats for another time).

Dave, who is a member here, but hasn't had much to say yet, has done a lot of experimenting and listening to get the L and C values in power supplies to just get the point where speed and stability meet. It might be there is a interesting point of comonality between whet the two of you are doing.

I guess, one day, someone here will have to try Jeff Medwins low DCR ideas (on AA) :-).
Whenever an honest man discovers that he's mistaken, he will either cease to be mistaken or he will cease to be honest.
Andrew
Eternally single
Posts: 4206
Joined: Thu May 24, 2007 2:18 pm

#5

Post by Andrew »

Nick wrote:
Interesting, have you any more details on that, power supplys are one of my areas of interest, Andrew and myself have had some interesting (sometimes beer fulled) discussions on them (including the prime number theory, but thats for another time).
Ah, I'd forgotten about the prime number idea! All that beer....

I have to say I think your multi-dimensional cap is an even better idea. I must see about building one sometime.

Nice looking amps btw, AndrewR, mine always look so....unfinished.....

cheers,

-- Andrew
Lee S
Old Hand
Posts: 664
Joined: Fri Aug 10, 2007 9:39 am
Location: UK

#6

Post by Lee S »

pre65 wrote:I thought of doing my PT15 monoblocks "mirror image" but it was too much work (for me).
Errrr.... Doesn't it just mean turning the top plate of one chassis upside down before fitting all the components, or am I missing something?? :wink:
©2020 Lee
User avatar
pre65
Amstrad Tower of Power
Posts: 21373
Joined: Wed Aug 22, 2007 11:13 pm
Location: North Essex/Suffolk border.

#7

Post by pre65 »

Hi Lee-not that simple !

I used a couple of "subframes" underneath on each one and then the wiring on tag boards has to be altered etc,etc.

I suppose if everything was attached to the top plate it might be less of a problem.

Have you tried it ?
The only thing necessary for the triumph of evil is for good men to do nothing.

Edmund Burke

G-Popz THE easy listening connoisseur. (Philip)
AndrewR
User
Posts: 44
Joined: Tue Oct 02, 2007 2:00 pm

#8

Post by AndrewR »

Nick wrote:Hi, nice looking amps,
Basically it uses descending values of supply caps, which works a treat!
Interesting, have you any more details on that, power supplys are one of my areas of interest, Andrew and myself have had some interesting (sometimes beer fulled) discussions on them (including the prime number theory, but thats for another time).

Dave, who is a member here, but hasn't had much to say yet, has done a lot of experimenting and listening to get the L and C values in power supplies to just get the point where speed and stability meet. It might be there is a interesting point of comonality between whet the two of you are doing.

I guess, one day, someone here will have to try Jeff Medwins low DCR ideas (on AA) :-).
Cool, what I have found is:
1) in a CLCRC only the last two C's need to be built with IDENTICAL sets of parallel capacitors
2) Of the set of parallel caps the highest cap must be quite high, in order to get an adequate level of smoothing. I started with 100uF on each, but it wasn't enough to get rid of the power supply a.c. components that was being amplified by the high level of gain in the ECC99 input and driver stages stage. So I went for 220uF as the highest value and virtually got rid of the hum. If I were to try it next time I would go for 470uF.
3) Next start choosing the descending values of parallel capacitors. My theory is this - the lower the capacitor's value the faster it will recharge and more responsive the PSU will be to transients. Try choosing the lowest value first - its effects are most obvious, and will give more agile treble and an "aura of speed" to the sound. Values should be chosen by ear. I chose 0.68uF LCR polypropylenes.
4) Now to choose the next value up. This value acknowledges that although the lowest parallel cap has a quick recharge response, it can only store limited charge. I went for 2.2uF, and it seemed to lend more speed to the lower treble. Again polypropylenes from LCR.
5) Now continue to add increasing values. I ended up with 4.7uF affecting the mid, and 10uF affecting the upper bass.
6) So for each of the last two C's in the CLCRC I end up with 220uF//30uF//10uF//4.7uF//2.2uF//0.68uF
7) I am finding that BHC capacitors are awesome, much better than Mundorf. The give a fuller underpinning sound (particularly bold bass) with a much more musical mid. They're great for the power supply and the cathode resistor bypass. They also take no-time at all to warm up - far better than Black Gate in that respect.
8 ) For whatever reason, I am finding that the Ducatis do not mix well with the amp, so I will remove them and replace to 10uF Ducatis with 10uF LCRs.

This is quite an adventure, and the I've learnt such a lot from the experience. All I can say is have a go, use some sets of crocodile clips and clip the pair of 0.68uF in parallel with the last two C's of your power supply, then try the other values, testing with your ears.

Also with regards to the choice of values I've also been wondering about whether sequences of values are best. I think the optimum would be along the lines of the Golden Ratio. However the range of values would probably relate as well to the frequency band and CR time factor, where R is the resistance of the amp as seen by the power supply and C is the resulting cap giving adequate voltage/charge recovery for its given area of operating frequencies.

Sorry if this is going on a bit, but to simplify the concept, it's like an engine where you have smaller pistons moving fast, and increasingly larger pistons moving increasingly slower and with increasing power.

Would be great to hear if anyone else has similar success in their projects.

Andrew Randle
Andrew
Eternally single
Posts: 4206
Joined: Thu May 24, 2007 2:18 pm

#9

Post by Andrew »

Hi Andrew

Sounds like you've been busy comparing lots of caps and combos. Some of the text books I have recommend bypassing large caps at either 1/100 or 1/10 of the value to get the effect you describe; I did this in my 300B; 470u, 4u7 and 470n.

Nick's and my beer fuelled discussion about prime numbers and I think your notion of the golden ratios is, I guess, about avoiding any one cap dominating over another at any particular load frequency; although there are a hundred other reasons why this might be a "good thing".

Nick's ultimate idea was to build a cap out of a arrays of smaller caps so that the capacitance would be high enough but the smaller caps would have better freq, and even, response; I like this idea.

Ultimately, I think that all passive PSU have their faults, and active PSUs are the way forward, tho' that's not the whole story either as its seems a little bit of RC afterwards brings back that midrange "love".

Currently, I'm trying to make a cheap and good solid state reg for some phono's I've been working on.

cheers,

-- Andrew
AndrewR
User
Posts: 44
Joined: Tue Oct 02, 2007 2:00 pm

#10

Post by AndrewR »

Hi Andrew
Andrew wrote:
Nick's and my beer fuelled discussion about prime numbers and I think your notion of the golden ratios is, I guess, about avoiding any one cap dominating over another at any particular load frequency; although there are a hundred other reasons why this might be a "good thing".
In a way, yes. However the way I look at it is that the caps operate over a limited set of frequencies, and I want the effect to cover the whole band without any gaps.

Ultimately the ears will lead the way, but something like the Golden Ratio may help.
Nick's ultimate idea was to build a cap out of a arrays of smaller caps so that the capacitance would be high enough but the smaller caps would have better freq, and even, response; I like this idea.
Sounds an interesting approach. I would the interesting to try two different arrays. One with a lot of the same small value (possibly bypassed with by a couple of even smaller values) vs. a set of descending values (which would be cheaper and may have more effect over a wider bandwidth).
Ultimately, I think that all passive PSU have their faults, and active PSUs are the way forward, tho' that's not the whole story either as its seems a little bit of RC afterwards brings back that midrange "love".

Currently, I'm trying to make a cheap and good solid state reg for some phono's I've been working on.
Cool, make sure you can test it both in and out of circuit. I'm not sure about regulators and wonder whether they are detriment to the "freshness" of the sound - but that at the moment is an unsubstantiated hypothesis and I may be wrong.

Thanks,

Andrew
AndrewR
User
Posts: 44
Joined: Tue Oct 02, 2007 2:00 pm

#11

Post by AndrewR »

By the way, I'm also looking at applying the power supply cap principles to my next project - a gainclone.

Not sure about the range of values and whether it would be the same. Will also have a look at using the Wima MKS2 & MKS4 range (metalised polyester) because they are smaller - anyone with any experience of this range of caps?

Best,

Andrew
User avatar
Nick
Site Admin
Posts: 15711
Joined: Sun May 06, 2007 10:20 am
Location: West Yorkshire

#12

Post by Nick »

One of the things I was trying to achieve with the large array of small caps, was to avoid one of the side effects of using diffiring values, I suspected that there must be a xover type situation gaoing on, the path to ground seen by a 100hz signal will be different to that seen by a 10kHz one.

Of course, thats not to say the idea is right, but it is a idea :-)
Whenever an honest man discovers that he's mistaken, he will either cease to be mistaken or he will cease to be honest.
JamesD
Old Hand
Posts: 997
Joined: Tue Oct 09, 2007 4:26 pm
Location: North Yorkshire

#13

Post by JamesD »

Just a quick note on using metallised polyester - they have a fairly high parasitic inductance that causes the cap to resonate in the 100+MHz region... my experience is to make sure no rf in that region is near the amp and make sure that gain is well below unity (>30db down) for that range when using them. Also if used in combinations make sure the inductance is snubbed or neutralised by design.

If you think valve audio amps won't operate at these frequencies then please don't ever try a 6C45pi or similar. Oh and typical valve inductors are a short circuit at these frequencies too :twisted:

ciao

James
User avatar
Nick
Site Admin
Posts: 15711
Joined: Sun May 06, 2007 10:20 am
Location: West Yorkshire

#14

Post by Nick »

resonate in the 100+MHz region
Thats usefull information James, one possible implication of that I can think of is the field from a nearby transformer could well contain big switching spikes from any rectification, this could (I guess) punp the cap into ringing even if its got ferrites on its inputs.

I can't help thinking some of the less positive results I have been finding using DC supplies on indirectly heated valve is due to this sort of noise getting into the system.

What happened to valve amps being simple to make :-( Its a bit like a fractal, the closer you look the more you find.
Whenever an honest man discovers that he's mistaken, he will either cease to be mistaken or he will cease to be honest.
AndrewR
User
Posts: 44
Joined: Tue Oct 02, 2007 2:00 pm

#15

Post by AndrewR »

Some interesting issues about the metalised polyesters. Fortunately in the gainclone, the PSU will be in a separate box with 22,000uF from each rail to ground. The main amps units will be in their own aluminium boxes :)

The only concern will be whether the parasitic inductance would be enough to smear any timing, if it doesn't kick-in until than 100MHz+ then it should be fine.

Best,

Andrew Randle
Post Reply