Hi, I'm Troy McClure. You might remember me from such self-help videos as "Smoke Yourself Thin", and "Get Confident, Stupid.". 😀
Recently I built two WE91A clones with re-winded WE-cores and mostly NOS components. I'm quite happy with the sound but wondering if it is possible to improve the high frequency roll off which it at around 11 kHz in my case. Schematic is very close to the original one with just one input stage. OPT primary is 3K.
I found a 91A report in an old VTV magazine which says that the roll off frequency is around 17 kHz for the 171A OPT.
Also ThorstenL (don't know if he's still online here) said that the DIY hifi-supply clones use a special trick to get this improved but I didn't find anything inside their schematics.
Any ideas?
He also was telling several times about the Ultrapath cap (or WE-Connection) which I'm using in my circuit but can't see in the original WE91A/B schematic. 😕
Regards,
Troy
Recently I built two WE91A clones with re-winded WE-cores and mostly NOS components. I'm quite happy with the sound but wondering if it is possible to improve the high frequency roll off which it at around 11 kHz in my case. Schematic is very close to the original one with just one input stage. OPT primary is 3K.
I found a 91A report in an old VTV magazine which says that the roll off frequency is around 17 kHz for the 171A OPT.
Also ThorstenL (don't know if he's still online here) said that the DIY hifi-supply clones use a special trick to get this improved but I didn't find anything inside their schematics.
Any ideas?
He also was telling several times about the Ultrapath cap (or WE-Connection) which I'm using in my circuit but can't see in the original WE91A/B schematic. 😕
Regards,
Troy
Attachments
Its the input pentode, the 310? is made on a pinch, the pinch produces big inter electrode capacitances . those capacitances roll of the frequency response.
A possible solution would be to take some of the bases used on a 310 (from junker tubes) and stick a EF86 plus associated circuitry on it to sub it for the 310A without modifying the amp.
cheers, V4lve
A possible solution would be to take some of the bases used on a 310 (from junker tubes) and stick a EF86 plus associated circuitry on it to sub it for the 310A without modifying the amp.
cheers, V4lve
I don't think it's only the 310A since the LD91 SET Amp is basically the same circuit (Frequency response: 12Hz to 47KHz):
LD91 SET Mono Block Amp | DIY HiFi Supply
I also tested my amp with other OPT without extreme HF roll off so I guess it has to do with the 171A behavior.
Cheers,
Troy
Here 'tis
Hi Troy - I remember you from such public information films as "Lead Paint - delicious but deadly ! "
See this thread, it's all in there :
Oh no - YAWE91B300BA (Yet Another WE 91B 300B Amplifier) - Thorsten - Tube DIY Asylum
I built it around 6/7 years ago , interesting project .
The two 1n caps back from the 300B to the 310A G2 provide positive feedback in the HF .
I found it was unstable in the HF at first switch-on, and had to decrease the value of the caps to around half . Do not just cut out one cap, or use one smaller cap , as the phase will not be correct !
Have fun !
( if you don't ultimately find the tone/body of the amp good enough, move on to a pentode with greater current, like an EL84, or a C3m )
Hi Troy - I remember you from such public information films as "Lead Paint - delicious but deadly ! "
See this thread, it's all in there :
Oh no - YAWE91B300BA (Yet Another WE 91B 300B Amplifier) - Thorsten - Tube DIY Asylum
I built it around 6/7 years ago , interesting project .
The two 1n caps back from the 300B to the 310A G2 provide positive feedback in the HF .
I found it was unstable in the HF at first switch-on, and had to decrease the value of the caps to around half . Do not just cut out one cap, or use one smaller cap , as the phase will not be correct !
Have fun !
( if you don't ultimately find the tone/body of the amp good enough, move on to a pentode with greater current, like an EL84, or a C3m )
Looks like you are using an Ampex input transformer. Did you try bypassing it to see if that's the culprit?
--------------------------------------------------------------------------
Below is the original schematic of the WE 91 for reference. It uses two 310 input tubes and has feedback going back to the screen-grid. Personally, I always feel that typical clone circuits that use only one input tube and minus the feedback are NOT the 91 circuit, regardless how much you like the sound. Anytime you delete an entire gain stage INSIDE the feedback loop, you're building a different circuit. I understand this is the result of the anti-feedback climate the audiophile community is living in.
---------------------------------------------
If you're a purist on WE diagram:
.
Last edited:
91 clone
Hi Troy, I have built three versions of this amp, including Thorsten's, and all have sounded good. I think the positive feedback Mark mentioned (btw Mark, I enjoy following your posts on the audiotalk site) would require you to have the additional gain stage of the original amp. I still have, and use two of these amps, one is the "daily driver". I am using a 6sj7 (5693) with the screen set by a vr tube, and am using three db of "schade" feedback by placing a 470k resistor from the plate of the 300b to the plate of the 6sj7. I measured the response of the amp at 1 watt, and it showed very little roll-off at 18-20k. The other 91 clone uses the same driver and feedback arrangement, but is operated at 2a3 op points, a simple voltage divider of the screen, and the feedback network used in the grid of the 300b. I am in the process of changing the filaments to dc on this amp, but even with ac, it only had about 2 millivolts hum at speaker terminals. I like the sound of these amps, so I really am not too worried about any of their short comings.
Hi Troy, I have built three versions of this amp, including Thorsten's, and all have sounded good. I think the positive feedback Mark mentioned (btw Mark, I enjoy following your posts on the audiotalk site) would require you to have the additional gain stage of the original amp. I still have, and use two of these amps, one is the "daily driver". I am using a 6sj7 (5693) with the screen set by a vr tube, and am using three db of "schade" feedback by placing a 470k resistor from the plate of the 300b to the plate of the 6sj7. I measured the response of the amp at 1 watt, and it showed very little roll-off at 18-20k. The other 91 clone uses the same driver and feedback arrangement, but is operated at 2a3 op points, a simple voltage divider of the screen, and the feedback network used in the grid of the 300b. I am in the process of changing the filaments to dc on this amp, but even with ac, it only had about 2 millivolts hum at speaker terminals. I like the sound of these amps, so I really am not too worried about any of their short comings.
That's exactly what I was looking for!😀
Found also this former post from Thorsten:
Re: thorsten we91 HF feedback loop problem - Thorsten - Tube DIY Asylum
I'll give it a try and adjust the values correctly.
Thanks a lot Mark! 🙂
@directdriver:
Thanks for the schematics.
Basically you are right but practically almost all so-called "WE91 clones and replicas" just have one 310A input stage. I have both schematics here for reference but for Hifi you just need one input stage. I bought this Kit ten years ago for a good price at ebay and all holes where drilled so I didn't want to change it. My personal definition for a WE91 clone is 310A(one or two)/300B/5U4(or 274).
@JimW:
Mark was right. I saw this caps some time ago but didn't recognize them as feedback. I also works with just one stage.
I prefer AC filament over DC if hum in not an issue. If you want to try it I suggest to use Rods version.
Kind regard,
Troy
There are exceptions such as the Shindo 300B amp. In USA, the single 310 tube circuit was popularized by Sound Practices magazine in the 90's, I believe. Without feedback, you do get enough gain in a single tube to drive a 300B (I am sympathetic to the chassis issue you're referring to as a DIYer) but I'm not sure that's the sound of the original. I've never heard an original 91 so I have no idea of its sonic character.
An externally hosted image should be here but it was not working when we last tested it.
An externally hosted image should be here but it was not working when we last tested it.
This article for Sound Practices magazine (1992) was my manual, so to speak😉.
I guess it brought the 'WE91 hype' to Europe and the US since before only some Japanese guys listened with very old cinema equipment from WE. All this stuff went to Japan in the 50s and 60s. Complete cinema tupe amps were sold for 1$ or thrown away and replaced by 'modern' transistor amps with lots of power.
I would love to here an original one but hear in Germany I guess its close to impossible to find one. The last one I've seen was sold for more than 50k!
I'm also not 100% sure where this typical sound comes from but I believe you can come quite close with just one 310A since the most critical parts for good sound are
1. OPT 2. 300B 3. Input tube
-> I think it's the output transformer... and the schematic/parts.
Cheers,
Troy
One significant feature of this circuit is that the 310A driver for the 300B is only running 2 or 3ma, which means it slews into the substantial miller capacitance of the 300B grid and rolls off noticeably by maybe 2dB at 20kHz. It definitely has a 'vintage' sound when run without the HF ( positive ) feedback - I think this is its appeal , it works very well with certain types of music .
I ran it with beautiful STC 4328D's as sub for the 310A's .
I keep experimenting with 300B drivers myself and keep returning to pentodes . Currently as of last night I'm running 4P1L pentode ( DHP ) at about 30ma . This is the best yet .
I ran it with beautiful STC 4328D's as sub for the 310A's .
I keep experimenting with 300B drivers myself and keep returning to pentodes . Currently as of last night I'm running 4P1L pentode ( DHP ) at about 30ma . This is the best yet .
I would be happy with 2dB at 20kHz but unfortunately the 171A transformers are acting more extreme in this circuit than other OPTs. 😕
I have several silver micas here (150p to 9n) and will look for the best compromise (positive feedback).
The STC 4328D looks like a 10J12S. Is it made in Russia or the UK?
For 310A I can choose between 10J12S, WE310A and WE328A. I also have an old Valvo 328A from 1940 with finest mesh anode but unfortunately only one...
My current 300B reference is the Russian NEM A300SE. It has an interstage transformer and a 6ZH43P (Double anode HF pentode) also running with 30mA. The schematic is as simple as it can be and without any cap in the signal path (battery for input bias).
Cheers,
Troy
I have several silver micas here (150p to 9n) and will look for the best compromise (positive feedback).
The STC 4328D looks like a 10J12S. Is it made in Russia or the UK?
For 310A I can choose between 10J12S, WE310A and WE328A. I also have an old Valvo 328A from 1940 with finest mesh anode but unfortunately only one...
My current 300B reference is the Russian NEM A300SE. It has an interstage transformer and a 6ZH43P (Double anode HF pentode) also running with 30mA. The schematic is as simple as it can be and without any cap in the signal path (battery for input bias).
Cheers,
Troy
Nice !
Some excellent options you have there.
6Z43P(E) is a new one on me - thanks for the tip-off .
The 4328D's were proper STC ones made in London .
Very nicely made, but not mesh plates .
Some excellent options you have there.
6Z43P(E) is a new one on me - thanks for the tip-off .
The 4328D's were proper STC ones made in London .
Very nicely made, but not mesh plates .
Sorry for the late response but in the middle of work my PC for measurements broke down and I had to buy a new (used) one.🙁
But now everything is working fine again and I was also able to improve my upper frequency from 11k to around 30kHz!🙂
I had to play around with different values and at the beginning I didn't get it stable. In the end a combination with a proper resistor value did the trick and I ended up with 1n-1n-18k in series (see frequency plot: upper curve 'new', lower curve 'old').
On the second frequency plot you can see that the harmonic distortion (2nd and 3rd) rises towards high frequencies but it's still in range and for the 3rd harmonic it became even more linear.
And yes, the 6Z43P(E) are quite unknown and you can just find a Russian datasheet. But in combination with an interstage transformer this small 0.8 Euro tube became my reference 300B amp... 😀
But now everything is working fine again and I was also able to improve my upper frequency from 11k to around 30kHz!🙂
I had to play around with different values and at the beginning I didn't get it stable. In the end a combination with a proper resistor value did the trick and I ended up with 1n-1n-18k in series (see frequency plot: upper curve 'new', lower curve 'old').
On the second frequency plot you can see that the harmonic distortion (2nd and 3rd) rises towards high frequencies but it's still in range and for the 3rd harmonic it became even more linear.
And yes, the 6Z43P(E) are quite unknown and you can just find a Russian datasheet. But in combination with an interstage transformer this small 0.8 Euro tube became my reference 300B amp... 😀
Attachments
- Status
- Not open for further replies.