• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.



to answer your question: the best and lowest-noise stuff available as new tubes seems to be the Sovtek 6922. Allen Wright uses them throughout in his preamps and reports only the best.
Forgetabout vintage NOS stuff, famous E88CC from Telefunken, Philips, Amperex and the even more fancy CCa from Siemens. They all sound slightly better than the Sovtek but are vastly inferior as far as long term noise behaviour is concerned.

Sorry in advance for my naysaying :)

You are bold to try that, MC step-up, do you know that?
ECC88 and family will soon turn from a new tube into a noisy tube.
Tube wizards having forgotten more that i ever will know failed in doing unless they paralleled ECC88 section, many of them.
I abandoned this plan and decided to use SS devices or better, input transformers.
Jensen JT-346-AX to be specific.

Point is: tubes have both thermal noise and shot noise and the shot noise is going to drive you crazy. S/N ratio of single tube sections is too bad for MC signal level: the signal partially drowns in the noise floor. If you parallel sections, this gets a bit better but you sacrifice detail resolution, so much that you better switch to MM cartridges. Reason for the MC cartridge is detail resolution and tone colour distinctability and saturation and exactly that you sacrifice by paralleling sections.

I have some planar ceramic triodes in my collection, G_m about 12mA/V, µ about 115. If i have much time surplus, i will try if the do the MC job. But i seriously doubt it.
Thank you for your sugestion regarding choice of the tube and certanly for sugestion regarding MC step up solution. I tried Lundahl 9026 and Ortofon TO-20 transformers, old PS audio solid state MC preamp (sounded the best), sort of Hiraga`s mc stepup with unsuficient powersuply..., and I was not satisfate. Combination of those devices with my Paravichini`s diy EAR835 riaa and my Ortofon mc25 on Kuzma turntable, simse to fall in reproduction. I was hapier with my old Grado entry level prestige on the same setup.
I`m shure that you are familiar with the design but anyway, this is what I`m trying to build (see Hampton schematic). It does use paralleled sections, design is extremly simple and what I was able to found about this preamp was encourageing. I`m aware of the problems I`ll hawe with the noise but I`ll take a risk and try.
Thank`s again!
rgds Marijan


  • hampton_mc.gif
    7.4 KB · Views: 2,909


2001-12-22 10:58 am
Boy did I see this schematic before?!

Marjan, If I'm not mistaken it uses a 200V B+?

I had this Hampton MC amp bookmarked but lost it last month. I think the web URL somehow did not serve anymore.

I have been using a similarly designed MC pre for the past 2+ years. It has a B+ of 24V. paralelled TFK 6DJ8's. The result has been satisfactory. With my MC Gold 0.4mV output, the loading resistors are 110 R. Noise has never been an issue. Neither was tube life.

Some say the 24V B+ is too low and not optimal for second order harmonic distortion and I should use 55V - I haven't had a chance to try yet. Hampton's schematic is worthy for a try and if you have a variac perhaps you would adjust to a 55V B+ and let us know your results.

This is original hampton suply. Basicly the same, changed in the heater dc section. B+ is the same 84V as on modification. I`ll try it on friend`s variac anyway. TFK diamond or normal?
BTW I have few toroids at home.What do you think about using separated transformers for plate and heater Planing to use 100V/150VA and 12V/60VA
rgds Marijan


  • hamptonsup.zip
    14.9 KB · Views: 359


2001-12-22 10:58 am
The new heater supply uses a LM317/LT317? That should be better than the 7806 alone. Don't forget to heatsink it.

I would use Scottkies diodes in both B+ and heater supply.

For "experiments" with a variac you have to use 2 transformers. Good luck and let us know your results.



diyAudio Senior Member
2002-08-21 1:20 am
MC Stepup using 6dj8

Sorry I spotted your problem this late.
I use a circuit that may just do the trick.
Basically I don't like to use transformers at this very critical frontend since any flaws will be amplified througout the audio chain.
Here is what you can do :

- 2 x 6dj8 in parallel per channel (dual mono) at 24 vdc regulated with a 7824 ic.
- regulate the 12 vdc heater supply with a 7812 ( 2 x 6 vdc)

- use a fet such as BF 244A as a current source for the anode current regulation ( or similar fet, idss must be the same ).

-cathodes are grounded.

-signal is taken from the anodes.

I used MIT metalized polysterenes since I could not find any colourations.(Metalized teflon is best on paper here but was not commercially available at the time)
Keep in mind to keep the signal path as short as feasable.

This sounds very well,noise is no problem unless perhaps if you'd
use ridiculously low output MC's.
Low level information is excellent provided the rest of the system is up to it.Transients are lightning fast,I use OTL amps so nothing there in the way of ringing or overshoot.
Frequency response is deliberately limited from 16 Hz to 1MHz in order not to burn the woofer coils.

The cartridge is not loaded and no input cap is used for ultimate transparancy.I used a Jan Allaerts MC hand built cartridge but others may require loading.Anyway that's simple enough to solve.

Best rgds,

Frank De Grove.


2001-12-22 10:58 am

The schematic I have was given by a friend and I had it for years. Anyway I think its not really complicated to build - routing of power supply hookups and grounding would be important in deciding the quietness of the amp.

Mine uses a 2SK30A ( 'GR" is said to sound best but I can only find 'BK') as current regulator. Other areas are identical except the ouput coupling has a total cap value of 10u and a 1M ohm output loading resistor.

The 24V was provided by a 25V EI transformer/scottkies bridge and a single LT317 adjusted to the voltage.

How would a cartridge work 'without a loading resistor'?

TFK 6dj8 (<>)has been my choice. I have tried a pair of sovtek 6922 and 6H23p and don't like them personally.
we are talking about the basicly same aproach to design here. Basic circuit is the same, whatever you chouse - one paralel or two twin triodes in paralel chain (less noise but maybe litle les informations?), biger or smaler cap, with or without output louding resistor. Heating - well, I decided to use 317 in TO-3 pack. (reason - realibility mostly) but the original version with 7812 and 7806 should do the job just fine. Basic diference in aproach is betwin original (see zipp file), or modified (see picture) version of regulation for plate +V and curent and the version you are talking about. 84 against 24V+.
Can you present me the arguments against original Hampton PS driving a tube on 84V and arguments in favuore of 24V aproach.
I allready ordered the parts for priginal version (mje340 instade of bd340, tl071 op.amp., tip122, mkp wima polipropilens...), but I would like to know the benefit`s of lower plate woltage regarding such a design.
Without the loading resistor? Well the lowest is 12ohm, what would 0 do - we should try, I have no expirience (never heard phono cube - 47laboratory)
BTW i have one Red label Valvo e88cc (gold pins), I`d like to have one more. If someone from Europa can suply me, please let me know.
rgds Marijan

did you consider to use the ECC88 as a choke-loaded common-cathode stage?
Provided the signal level is not too high (the tube usually had -1.3 V bias only)
this could work well, but then you have gain 33. For a juicy linestage, i would rather choose a 5687 and an step-down output transformator. Like the pa102 from Simon Shilton. Or a 12B4A ....

dunno whetrher the ECC88 is the best choice as MC input tube. You need low noise, so use a tube with super-high transconductance, 5842 is probably very good, G_m twice of ECC88. And it is more linear, this being not so important in an MC headamp. Or you try what i try: EC8010. µ=60, G_m=28mA/V, even higher than the 5842. OKok, there is this EC8020 with µ=55 and G_m=60, it is sonically gorgeous but completely unobtainable and unaffordable.... And Thomas Mayer who uses EC8020 still uses an MC input trannie for noise reasons.
HE uses the same as i do: the Jensen JT-346-AX. Nothing else for me please.
I know the Lundahl trannie, it is very nice but not as good as the Jensen.

i did not adress the option to use ECC88 with that low B+. Nice idea! If this works fine, i stand corrected :) Guess i have to try that out one day.


diyAudio Senior Member
2002-08-21 1:20 am
Hello Bernard,Marjan,et all,

Regarding the MC headamp using the paralled 6dj8:

-the tube was chosen since it is suited for use at vg = 0.
Since we're talking microvolts here overload is no problem.In reality it is actually amazing how much inputswing this tube can take.
I actually tried it as a line stage and it still refused to overload,I'm not saying the circuit is usable in that application,just to give you some idea.
-The important factor in the design is the high transconductance of the 6dj8 here.If you want lower noise try 7308/E188cc if you can find four decent ones.Personnally I can't be bothered,we choose the tube since it's readily available which allows careful selection.
**The major difference with the Hampton circuit is the use of a constant current source BF244A wich provides the plate with a constant 2mA current (IDss).Other seniconductors are available but this one is around in abundance.I stress it must be the A version B and C version won't work.
The output resistor is usually set to 100k since in normal use the user will have a MM stage following the headamp.
Since the headamp sits in front of any RIAA corrected stage one can use a relatively small coupling cap to the next stage say 0.68 micro F.
Smaller values usually sound better then big ones.
Other tubes are available from Russia now wich are even better candidates for this appication,but I would need to trace their curves for that.Anyone ?
Actually the cartridge loading resistor on a mc cartridge serves as a break on the coils.You should try the effect of doing away with it in your system.I'm aware that results will vary.
My design philosophy is of the Zen approach , meaning less is more.Less components.less gain stages and so on.
Since we're talking DIY here it is easy enough to design a phono +
line stage in a mere three stages,four if you use passive RIAA correction (high insertion loss )
At the end of the day it all depends on the efficiency of your speakers and other lay out requirements such as cable lengths etc.

John, Bernhard
thank`s for your advices regarding the tubes choice. Today I picked up a pair of NOS Tesla ecc88`s I`ll try them first.
Frank for your view
I`d be glad if you can send a schematic of your version of MC modul or at least PS. Since you are using MC step up based on a similar design but with the diff. voltage I`d like to see changes comparing to the Hampton.
rgds Marijan


diyAudio Senior Member
2002-08-21 1:20 am

Unfortunately I don't have the schematics of this circuit at hand.
Still the beauty of it lies in its simplicity.
All you require for it to work is a handful components:

1/ 1 x Tr with 2*30 VAC sec 100 mA/ea is more than enough.
8 x 1N4003 diodes or their Schottky equivalent arranged in a full wave bridge.
2 x 50 Vdc caps for ac ripple filtering.I used 4700 micro F.
2 x 7824 IC with their housing/ground srewed down to the metal housing for cooling (use heat conducting paste )
2 x coupling caps of good quality and at least 50 V insulation.
( generally higher voltage caps sound better) .Anything starting from 0.47 miro F will do.
2 x loading resistor (Vishay bulkfoils are excellent but expensive) for the cartridge you use.
2 x output resistors 100 K Holco 1/4 W

**you can shunt the 7824 with a small polyprop cap but look at the datasheet for max. value and put it just before the FET current source.

2 x BF244A Fet : source connected to B+ and Gate and Drain connected to V1-2 anode.Remember Idss no more than 2 mA.

2/ 1 x Tr with 2*15 VAC 800mA/ea see 1/ for filtering and use 7812 IC.

3/ Use star grounding throughout and route heater and B+
voltages carefully as to not pick up interference.

4/ Connect internal shield of the 6DJ8's to the common grounding point.

** Anything more fancy I let to your imagination but will be on the expense of complexity.
I use hardwiring with soldertags and solid core hookup wire throughout.

Greetz and keep us posted,

thank you very much for your input. I need differential input but the stage you describe is feasible as long-tailed pair.
zero V_g and low B+: great idea, didn't muse about that. I will try it out and compare it to my Jensens.
In one respect i am stubborn: i do not parallel tubes. Moreover i have no use for gain 33 (ECC88 with CCS as load), gain 18 or 10 sounds more to my taste. I already have an idea how to integrate this thing DC-coupled into my preamp: long-tailed pair, plate resistors say about 2k (one fix, one adjustable), Grids floating at about -25V, long-tailed pair CCS tied to minus 50V supply. Plates tied to ground. I did not calculate this so far, so please do not nail me down on the voltages and values....

For a short descripition of my preamp, see 12B4 thread

i may have some NOS-looking Tesla E88CC surplus. Email me privatly.


diyAudio Senior Member
2002-08-21 1:20 am
Hello Bernhard,

As I read both your threads and also the one on your MM cartridge mods,here are some thoughts:

-Keeping all else equal CCS and B+ I think there is a good chance you may be bugged by residual tube/shot noise in the first stage.
Also in my opinion: cascading same topologies amplifies its shortcomings along with it.Try it anyway.

-The first stage can be used as a diff amp as well without going for the long tailed pair topology.This goes for other circuits as well.I keep on noticing everybody adopting the long tailed pair as the only way to go to make a differential stage,it is not.

-Since you intend to lower the gain of the first stage (and I don't know the gain of the next) you risk to amplify noise if your second stage has higher gain.

-Re MM :most manufacturers had the habit of tying one of the outputs to ground internally rendering its inherently balanced output single ended. (This was mainly done to avoid ground loops and hum pick-up.)