• 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.

Simple Hi-end 845 amplifier

Status
Not open for further replies.
Jump to Latest
#21
 
#23
What is the purpose of the 12AX7 in the front?
You don't need all that gain and you amplify noise too.
Everything with a gain about 12 (6SN7 with 47k load for example) is capable to drive the rest to full power with 1 Vrms input.
An 6SN7 is much more linear than your choice at about 12Vrms output (that is what you need from the first stage to achieve full power). And it's 8-10 k internal resitance is lower even than 4 paralleled 12AX7.
Strange choice indeed, or has this something to do with the input cable, which is so detailed specified? Should i assume that e.g. a different length will affect the "sound"? I am pretty sure that any other type than ..g-audio was that?... will not sound half as good

The EL34 has the interessting feature, that the distortin spectrum consists almost of 2nd. 3d is very low. Even if the total THD is about double as with a 2A3, 3d is much lower.
The cancelation with the output stage will produce a lower THD and 3d than with a DHT driver. If it sounds good, it's another story, but i would give the EL34 a try too.
Regards
Konstantinos

:whazzat::whazzat:
 
#24
About input stage:

It is true, the gain of input stage is too high for the necessity.

I have forgot last evening to test the sensibility because I was concentrate on distortion level and spectrum so I will do soon a new section of measurements.

About the internal input cable:

Offcourse the length 40cm is necessary only with my chassie layout but I would like to emphasize the importance to this components that give a very strong contribution to the final result.

Q-Audio has been my choice after my test on different cables (see section on my website)

P.S. Sorry to all for my bad english
 
#25
audiodesign said:
I am interested to know if someone have valid ideas to create a better design without use old and strange tubes.
Click to expand...
6N6 ( overkill), ECC88, 6SN7, even ECC82 would be a better choice (from a distortion and driving ability point of view) than this thing and equally important:
less gain---> less noise! In this case, you wouldn't need DC for the first 2 stages, especially with lowish -90db- sensivity speakers.

Greetings
Konstantinos

P.S. You don't need to pay huge amounts of money for the clio.
A good sound card -Maudio 192 to name one- is more than adequate for DIY. How must did it cost? or it didn't?
 
#26
schiller said:

6N6 ( overkill), ECC88, 6SN7, even ECC82 would be a better choice (from a distortion and driving ability point of view) than this thing and equally important:
less gain---> less noise! In this case, you wouldn't need DC for the first 2 stages, especially with lowish -90db- sensivity speakers.

Greetings
Konstantinos
Click to expand...

Yes, a nice ecc82 - it will sound quite much like a 6SN7 since the construction is similar (two 6J5 triodes in a glass envelope).

Changes would also be minimal too.

I would avoid dc heaters if at all possible. If it is too noisy think about a grid stopper. Then try the MJ Thingy. And only then try DC. Why? DC is not easily done well. When it is not done right it sounds like wet cardboard. You only really need it on phono anyway IMHO.

I believe Brett wants to see a nice linear response across the audible spectrum. Andrea insists on very low inherent distortion. They are not exactly the same objectives but they can be met.





:smash:
 
#27
I wouldn't place the ECC82 high on my list of recommendations, it's linearity is at least 10dB worse at a given amplitude than the 6J5, 6SN7 or 6FQ7/6CG7. I think SY and a number of others can confirm this, Morgan Jones writes at length about the linearity of various tubes out there, and the ECC82 wasn't nearly the best. IME in SRPP designs I have measured the 6SN7 was up to 12dB better in terms of overall THD at a given quiescent current and swing.

In terms of currently easily available tubes I would look at the ECC99, inexpensive Russian tubes like the 6S4P (very similar to 5842) - I paid about $4 each for a bunch of these on eBay just a couple of weeks ago.
 
#28
audiodesign said:
I am interested to know if someone have valid ideas to create a better design without use old and strange tubes.
Click to expand...

Not really. The "old tubes" originated at a time when the use of NFB was quite rare. For the best sonics, these types were designed to be quite linear for use with amps that ran open loop. If you don't apply corrective NFB, there is no other way to get good, low distortion sonics.

During the post-WW II era, gNFB came into wide -- and excessive -- use. This created the demand for high gain types like the 12AX7 so that high open loop gains could support 20+db(v) of gNFB and retain a useful sensitivity. Gain, not linearity, was the main design criterion. What difference does it make if 12AX7s aren't as linear as 6SL7s if you can simply add tons of gNFB to force linearity?

I think this is the wrong way to go, and that no VT amp should ever require anywhere near 20db(v) of NFB. If the design does need that much to linear up, then the open loop design is defective and needs to be redone. Obviously, this was not done and the end result was a whole bunch of sound-alike amps.
 
#29
audiodesign said:
About input stage:

It is true, the gain of input stage is too high for the necessity.

I have forgot last evening to test the sensibility because I was concentrate on distortion level and spectrum so I will do soon a new section of measurements.

About the internal input cable:

Offcourse the length 40cm is necessary only with my chassie layout but I would like to emphasize the importance to this components that give a very strong contribution to the final result.

Q-Audio has been my choice after my test on different cables (see section on my website)

P.S. Sorry to all for my bad english
Click to expand...

Hi, you don't have to measure gain, you can calculate it, even before you build !!
12AX7 has with your values in the circuit about 50, EL34 7-8 at this operating point, 50x7=350.

You need about 75 to drive the grid of the 845 to full output with 1Vrms input. You amplify noise 4,5 times more than necessary.
If you have noise problems (you may not, maybe your input is quiet), this is a possible source and not the AC filaments of the 2 IHT's, it is not a phono preamp.

About the cable, i knew it. It is obvious to anyone that 40cm wire can make or brake the perfomance of a circuit which consists of 3 non-linear active devices, 2 transformers e.t.c.

I believe it is the first time you make distortion measurements, this explains why you seem to be thrilled with the results. They are low values - in absolut terms - indeed, but just mediocre in comparison to a proper engineered input stage.
Try for exampe an ECC88 in the front and you will see it. Use e.g. 47k anode load, 4 mA anode current. Not the most linear op-point, but you need only to swing 10 Vrms, that's to easy.

And there is a simple test to find out if the higher flux in the IT is a problem, you don't have to speculate:

Drive to full output, measure distortion at the 845 grid at 1kHz. Repeat with lower frequency, 50,40,30,20 and note where dist. components start to rise (they will). Define a distortion level as acceptable (for me it is 0,5% additional 3d harm.). The lowest frequency where this is not exeeded is your limit for full power. If it is too high for you (for me again, 30Hz is the limit), yes you have a problem.
The hole procedure is less time consuming and much more usefull than testing input cables with the ears (IMHO). You can do it of cource, you live in a free country too, but at least optimize the circuit first, before you start hearing differencies between wires.

Best Regards
Konstantinos

P.S I think your claim at your site:
citation
So the result is a perfect amp. and I am sure there are very few 845 amp. around with this performances.
end citation
needs to be revised, let's be serious
 
#30
kevinkr said:
I wouldn't place the ECC82 high on my list of recommendations, it's linearity is at least 10dB worse at a given amplitude than the 6J5, 6SN7 or 6FQ7/6CG7. I think SY and a number of others can confirm this, Morgan Jones writes at length about the linearity of various tubes out there, and the ECC82 wasn't nearly the best. IME in SRPP designs I have measured the 6SN7 was up to 12dB better in terms of overall THD at a given quiescent current and swing.

In terms of currently easily available tubes I would look at the ECC99, inexpensive Russian tubes like the 6S4P (very similar to 5842) - I paid about $4 each for a bunch of these on eBay just a couple of weeks ago.
Click to expand...

I will have to search through those older threads... I have to agree that there are other alternatives, but as a drop in replacement you will notice a difference right away with an ecc82. Gain would be reduced but sound quality will be immediately improved.

I'm not sure if you could just drop an ecc99 right in there... You wouldn't want to do that with an ecc88...

I know this from building another of Andreas designs. I can confirm that most of his designs sound better than you would expect - that might have a lot to do with the iron used.

I will have to check out specs on the ecc99 now though. I do agree entirely with shiller's comment.
 
#32
About the input stage:

I am considering to try Ecc82 instead of 12ax7 to decrease the gain and check the sound quality.

but there are many configuration used by DIY where this high gain is necessary, for example if you want connect this amp. directly to a DAC chip with low voltage output like TDA1541, PCM63K, PCM1704, AD1865 used offen with low output resistor like passive I/V

I have seen a good TDA1541 kit on:
http://us.hifidiy.net/Article.asp?ArticleId=184
(escluding vaccum tube output stage)
http://us.hifidiy.net/Article.asp?ArticleId=185
(escluding opamp output stage)

About the comparation with OTL:

The OTL is another world ! The my OTL has been the first big project with vaccum tubes so it is very old.
It is quite impossible create a good OTL because feedback and vacuum tube used (normal regulator) decrease the final result.
 
#33
One of Chris, DIY - AVclub.gr member, recent 845 SE implementation.
Together with AE amorphous OPT sounds just great.
It's a similar design as previous diyaudio members thoughts.
 

Attachments

  • 845 se avclub_gr.jpg
    845 se avclub_gr.jpg
    38.3 KB · Views: 1,166
#34
The amorphous OPT should give better performances, I hope to try in the future the Lundahl with amorphous core.

I don't like the inducance, instead of transformer, like interstage in the driver section because add an extra capacitor.

If we sum the cost of inductance and a good cap. the price will be higher than tranf.

The extra cap. limit aslo the band if it is too little and there are some transation in class A2.

Probably a LL1692 with primaries in series will be good to be driven by 2A3 at 36mA (LL1692A/18mA 35H at 35mA)

I am thinking also to add an interstage in first stage.
 
#35
The OTL is another world ! The my OTL has been the first big project with vaccum tubes so it is very old.
It is quite impossible create a good OTL because feedback and vacuum tube used (normal regulator) decrease the final result.
Click to expand...

Thanks!

And what about your mosfet follower? You wrote: with a good tube preamp it can beat a 300B SE.

Tyimo
 
#36
I think this is not the right place to speak about my Power Follower (pf99 or pf2007) but in any case:

1) I know of some peolple that sell their 300b amp. and now they are using my pf99 with a good vacuum tube pre-ampl.
Many DIY with lot of experience with tubes can confirm that the 300b is not a valid tube, and I confirm it on many types tested like SV300B, 4300B, KR300B.
This because the 300b have a strange characteristic so not very neutral like 2A3 or 845.

2) the pf99 is an opportunity to get very hi-end sound with low cost.

3) the pf99 is secure, vacuum tubes sound with low voltage.

4) the pf99 is an opened system because you can change the inpu stage to get different sound from solid state solution with a single IRF610 to more complex vacuum tube totem pole or SE stage.

5) there are many other methods to create an hybrid amp. but the pf99 is the first choice if we consider the sound qualty.

6) In a recent comparation with this new 845 amp. the pf2007 have lost but I consider this as natural because a good pure vacuum tube amp. with high damping factor and with the best components on the market always will give a sound with more emotions.
 
#38
In a recent test on my 845 amp. ver. 1 I have compared the LL1620/90mA with the Tango X-10S.

There was some little differences but both give the same sound quality and it was impossible decide with is the best.

This confirm my idea about Lundahl, It produce good components at a good price (Tango 1250$ per pair - Lundahl 150-190$ each).

Have you good alternatives ?
 
#39
I would like to hear what you think about using the ecc82 as a driver in this amp.

I use the ecc82 (cv4003) in my build of your parallel single ended el34 - but only for classical music. For jazz, blues and pop I use an ecc83 (cv4004).

The gain is significanly less with the ecc82, but a change in tonal coloring becomes moderately apparent. Other valve choices would require re-wiring, and it is nice to just swap this valve.

My version of this amp is dead silent at full volume and I have well over 90db sensitive speakers. It is very musical and detailed with AC heaters.

In a way it sounds more like PP, except with the added detail of SE. I think that a part of this is actually due to the ecc83 in the gain section.

The ecc83 does not have a linear response across the audible spectrum. It is clearly stronger in the mid range. But it definitely not 'poo' or 'dog poo' as some posters like to suggest. It can be used to great effect depending on its application.

The e88cc, with a gain of 33 would be an even better (9-pin noval) choice for more linear response than the ecc82. The heaters will need to be rewired though.
 
Status
Not open for further replies.
Top Bottom