i build the Hiraga Optime Tarnsimpedance RIAA. Actually there was not much to build because i had already the RIAA part from another project. So i had only to insert C1. I found a mistake in my last drawing. C1 should have 100nF and not 10nF. It can be connected in two ways : to the positive supply or to ground. The connection to he supply had 0.5dB more gain but i decided to put the cap to ground because it was more compact to solder. All other measurements were the same. I used the LT1468 as opamp and a AD711 as servo. The servo will not do any harm because it is very slow. If in doubt you can add a low pass filter in series with the servo output like 10kOhm and 1uF.
The sound i got was good, fat and dynamic. After listening to two records i found it a bit on the analytic side. I usually like the LT1468 very much. It has a one stage folded cascode bootstrapped current mirror layout with a lot of gain so distortion from this OP is one of the lowest you can get. The sound can be slightly "matter of fact" and i missed the lush sound of the Telepath. I substituted for an OPA827 and most of the sound of the Telepath was suddenly there. More relaxed and liquid with very good differentiation of tonal colours. After some more records i still found the sound a bit "grey" so i changed the tone arm cable from a 1.5m run of Wire World Silver Eclipse 5.2 that had worked so well with the Hiraga Optime - Telepath to my Spiral Grove 0.5m chaos-fractal cable and suddenly the sound opened up to the wall to wall size i sometimes get if i have luck. tonal balance was restored too and there was much more resolution and micro detail without spittyness. My dear subjective friends, whatever the experts tell you : cable is very audible especially when you get closer to the tone arm. I have no hard scientific fact to offer but i have a feeling what happens. Anyway the sound is now 90% of the Telepath and has no right to be that good. the telepath is not cheep and slightly better but this can be build for 300,-€ including cabinet. Do not skimp
on the power supply. I use four power supplies in a "double - double - mono" configuration with coils in the supply lines and hope that Salas circuit sounds even better.
Some other tips : I use 47kOhm for R9. That gives a curve that looks like a straight line that rises by 0.5dB from 20kHz to 20Hz. I found that response best in my system subjectively. My Titan is not short on treble and my speakers are very extended on top and i sit close so you may opt for 220kOhm that will give a response that is + - 0.1dB. my measurement limit. I use Vishey/Röderstein MKP1837 1% and simple 1% metal films here. By coincidence C1 and C6 have the same value so you can buy a bunch and select as close as you can. 100nF is not an exotic value anyway so most Diy´ers should have some in their stock.
The sound i got was good, fat and dynamic. After listening to two records i found it a bit on the analytic side. I usually like the LT1468 very much. It has a one stage folded cascode bootstrapped current mirror layout with a lot of gain so distortion from this OP is one of the lowest you can get. The sound can be slightly "matter of fact" and i missed the lush sound of the Telepath. I substituted for an OPA827 and most of the sound of the Telepath was suddenly there. More relaxed and liquid with very good differentiation of tonal colours. After some more records i still found the sound a bit "grey" so i changed the tone arm cable from a 1.5m run of Wire World Silver Eclipse 5.2 that had worked so well with the Hiraga Optime - Telepath to my Spiral Grove 0.5m chaos-fractal cable and suddenly the sound opened up to the wall to wall size i sometimes get if i have luck. tonal balance was restored too and there was much more resolution and micro detail without spittyness. My dear subjective friends, whatever the experts tell you : cable is very audible especially when you get closer to the tone arm. I have no hard scientific fact to offer but i have a feeling what happens. Anyway the sound is now 90% of the Telepath and has no right to be that good. the telepath is not cheep and slightly better but this can be build for 300,-€ including cabinet. Do not skimp
on the power supply. I use four power supplies in a "double - double - mono" configuration with coils in the supply lines and hope that Salas circuit sounds even better.
Some other tips : I use 47kOhm for R9. That gives a curve that looks like a straight line that rises by 0.5dB from 20kHz to 20Hz. I found that response best in my system subjectively. My Titan is not short on treble and my speakers are very extended on top and i sit close so you may opt for 220kOhm that will give a response that is + - 0.1dB. my measurement limit. I use Vishey/Röderstein MKP1837 1% and simple 1% metal films here. By coincidence C1 and C6 have the same value so you can buy a bunch and select as close as you can. 100nF is not an exotic value anyway so most Diy´ers should have some in their stock.
After cooking my famous ginger chicken soup for my family i had more time for listening. I am preparing for Holgers visit so i looked what i could improve still. Cranking up the sytem full (that is so loud that it whould destroy my speakers) i still heard a faint hum. I found it came from my unshielded chaos, fractal cable. I build a shield from copper tape and braid i had ripped of a CX100 coaxial cable, connected the shields to the Hiraga earth and the hum was gone. The sound improved too : cleaner in the bass and more liquid in the midrange- treble. Strange that this short cable is so important. Sound is now very similiar to the Hiraga-Telepath combination with the exception of the stratosferic treble where the Telepath has the edje in resolution and speed. That can be explained in two ways : 1: i am using better parts in the Telepath like Panasonnic FM, Dale RS55 and Polystyrol capacitors for the RIAA, 2: the Telepath has a 2 stage RIAA and a very fast and strong output buffer. The poor OPA827 in the Transimpedance RIAA has to amplify 50dB in the bass and 30dB in the midrange treble plus driving a 5m long video coax cable. Still i like the treble performance very much. It´s sweet like liquorice and i could imagine listeners that prefer that over the Telepath.
Anyway, i will play this combination for Holger and switch later to the MPP low Z with inductive RIAA.
Holger, when you read this please bring Romans Preamp with you. Roman will borough it to me now. It´s a Vacuum State, so now i have a chance to listen to it in my home.
I have also more good news. i am dedicated to build a compleet DIY system with phonostage, preamp, poweramp and speakers, so stay tuned. I will try to use designs that are popular here like the B1 buffer and the Pavel Mosfet Buffer and will try to squeeze as much performance out of them as posibe.
Anyway, i will play this combination for Holger and switch later to the MPP low Z with inductive RIAA.
Holger, when you read this please bring Romans Preamp with you. Roman will borough it to me now. It´s a Vacuum State, so now i have a chance to listen to it in my home.
I have also more good news. i am dedicated to build a compleet DIY system with phonostage, preamp, poweramp and speakers, so stay tuned. I will try to use designs that are popular here like the B1 buffer and the Pavel Mosfet Buffer and will try to squeeze as much performance out of them as posibe.
Just for fun i designed a passive RIAA version of the Hiraga Optime Transimpedance.
It has the following advantages: !: Gain split equal and frequency independent over two stages. 2: Strong and fast output stage for driving long cables. 3: Some people like passive RIAA better.
OP 1 could be an LME49710, OP 2 = OPA134, OP 3 = LME49713 or AD844.
I took the closest values for cap and resistors i could find. The formulae are : R15 x C2 = 3180uS, R1 x C2 = 318uS. C2 is quite big and i would recommend Evox/Rifa PHE450 here. They come in 5% tolerance but are mostly better, more like 2%. It should be no problem to find well matched pairs.
It has the following advantages: !: Gain split equal and frequency independent over two stages. 2: Strong and fast output stage for driving long cables. 3: Some people like passive RIAA better.
OP 1 could be an LME49710, OP 2 = OPA134, OP 3 = LME49713 or AD844.
I took the closest values for cap and resistors i could find. The formulae are : R15 x C2 = 3180uS, R1 x C2 = 318uS. C2 is quite big and i would recommend Evox/Rifa PHE450 here. They come in 5% tolerance but are mostly better, more like 2%. It should be no problem to find well matched pairs.
Hi Joachim,
I have been off this forum for a while, when I spotted this thread yesterday.
I have been playing with various phono stages, and I am still trying diy solutions that give me all I need for exciting audio playback one day.
It would be a mix of the performance of a phonoclone, which, in my incarnation, give that certain slam, the interaction of bass and drums and rhythm guitar, that goes right down to your bones when heard live and, sometimes, on master tapes (or direct copies), but rarely on vinyl (let alone CD), but lacks transparency and resolution compared to a toshiba-fet input, no loop feedback design, that in turn doesn't (re?-)produce that certain slam.
I have been partly successful with a rambeyrolles input stage, but with loading of the folded cascode output instead of loop feedback, and and a no loop feedback second stage and passive RIAA. Still, the other stages are superior in their special department.
In the meantime, I use an odd one-stage design that sounds good in respect to complexity and price. No audio nirvana, though.
Your reports sound, well, sound, so I will eventually try some of your ideas. (a friend of mine runs audio physik speakers that sound great, that is a reference for me as well)
I looked again at your Hiraga Optime: you chose npn instead of pnp-devices. Maybe it slipped me, but do you have an explanation for this choice?
Thanks for this comprehensive thread!
Rüdiger
I have been off this forum for a while, when I spotted this thread yesterday.
I have been playing with various phono stages, and I am still trying diy solutions that give me all I need for exciting audio playback one day.
It would be a mix of the performance of a phonoclone, which, in my incarnation, give that certain slam, the interaction of bass and drums and rhythm guitar, that goes right down to your bones when heard live and, sometimes, on master tapes (or direct copies), but rarely on vinyl (let alone CD), but lacks transparency and resolution compared to a toshiba-fet input, no loop feedback design, that in turn doesn't (re?-)produce that certain slam.
I have been partly successful with a rambeyrolles input stage, but with loading of the folded cascode output instead of loop feedback, and and a no loop feedback second stage and passive RIAA. Still, the other stages are superior in their special department.
In the meantime, I use an odd one-stage design that sounds good in respect to complexity and price. No audio nirvana, though.
Your reports sound, well, sound, so I will eventually try some of your ideas. (a friend of mine runs audio physik speakers that sound great, that is a reference for me as well)
I looked again at your Hiraga Optime: you chose npn instead of pnp-devices. Maybe it slipped me, but do you have an explanation for this choice?
Thanks for this comprehensive thread!
Rüdiger
Hi Onvinyl !
Yes, the Rambeyrolles has some issues to solve. I build the Hiraga Optime with PNP but have shown the NPN circuit. They both work, just reverse everything. Some people like PNP better because of lower noise but i think with this optimised transistors it is no issue.
I looked at your one-stage-design and think it is not bad at all. What i would do is to use the output of the 2SJ74 too for amplification. You waiste a lot of potential here..
Yes, the Rambeyrolles has some issues to solve. I build the Hiraga Optime with PNP but have shown the NPN circuit. They both work, just reverse everything. Some people like PNP better because of lower noise but i think with this optimised transistors it is no issue.
I looked at your one-stage-design and think it is not bad at all. What i would do is to use the output of the 2SJ74 too for amplification. You waiste a lot of potential here..
Yes, as soon as I see an elegant way to do it while keeping it single stage.
🙂
Rüdiger
Joachim, Holgar,
If you are getting to hear a Vacuum State RTP3C/D, a word of advice.
Please do NOT listen to it until it's been on for AT LEAST 24 hours. For whatever reason, it sounds quite ugly until it has really warmed up - and once it's on, leave it on please. It's intended to run constantly, 24/7/365.
That does NOT burn out the tubes - in fact they last much longer if never turned off!
Second point: it's a conventional voltage input on the phonostage which has provision for loading Rs. I strongly suggest you use 47k, and nothing lower.
The RTP has enough input overload headroom as to not be affected by any ultrasonic cartridge resonances - and 47k loading allows the full dynanics of the cartridge to shine through.
Attached is a schema of the phono section, it feds a balanced selector and attenuator, and a simple/complex line stage.
If you send me your mailing address, I'll send you a Xmas present.
Very interested in your opinions.
Regards, Allen
If you are getting to hear a Vacuum State RTP3C/D, a word of advice.
Please do NOT listen to it until it's been on for AT LEAST 24 hours. For whatever reason, it sounds quite ugly until it has really warmed up - and once it's on, leave it on please. It's intended to run constantly, 24/7/365.
That does NOT burn out the tubes - in fact they last much longer if never turned off!
Second point: it's a conventional voltage input on the phonostage which has provision for loading Rs. I strongly suggest you use 47k, and nothing lower.
The RTP has enough input overload headroom as to not be affected by any ultrasonic cartridge resonances - and 47k loading allows the full dynanics of the cartridge to shine through.
Attached is a schema of the phono section, it feds a balanced selector and attenuator, and a simple/complex line stage.
If you send me your mailing address, I'll send you a Xmas present.
Very interested in your opinions.
Regards, Allen
Attachments
No pun intended Allen, but I would be really curious what are the advantages (or secret ingredients for performance) in this schematic. That's because by all engineering standards, it has poor performance (noise, input impedance, offset, etc...). I also miss the purpose and advantage of cascoding bipolars with tubes.
Was it designed following "subjective sound" criteria only? That would explain some of the decisions...
syn, you have me scratching my head. Noise is very good (0.8nV/rt Hz), the tubes give very large and linear swing, input impedance well high enough to not bother a cartridge. No output offset, it's cap-coupled. What's the objection?
Stu, I am scratching my head as well. For example, 2x22ohm in parallel with the input makes 11 ohm, or about 0.4nV/rtHz noise by itself. And this is before considering half of the emitter resistor.
I would understand tubes in a gain stage, but what's the advantage of a tube as cascode? A high voltage mosfet will do exactly the same, while the cascode noise doesn't matter anyway.
Using the SSM pairs, what's the current noise of this amp?
And why feeding the RIAA network from a finite impedance, while a (e.g.) folded cascode (like in the Vendetta) will provide a very high impedance?
Last edited:
Well, yeah, there's that. One more reason to use transformers at the input.
I haven't looked at the distortion spectra for MOSFETs on top- I have seen Allen's results with FETs on the bottom, tubes on top, and they're quite good.
Oh my, you'll have fun with my RIAA article. 😀
Transformers are terrific in this application, though I certainly understand the prejudice against them.
Transformers are terrific in this application, though I certainly understand the prejudice against them.
[QUOTE=syn08;
but I would be really curious what are the advantages (or secret ingredients for performance) in this schematic. That's because by all engineering standards, it has poor performance (noise, input impedance, offset, etc...). <<<
Can you please explain why it has poor noise, poor inpt impedance, poor offset - and why these, other than noise, and to a very small degree, offset, are important in a MC phonostage?
>>I also miss the purpose and advantage of cascoding bipolars with tubes.
Was it designed following "subjective sound" criteria only? That would explain some of the decisions...<<<
I'm happyto amswer your questions after I'm clear as toyour complaints.
Regards, Allen
but I would be really curious what are the advantages (or secret ingredients for performance) in this schematic. That's because by all engineering standards, it has poor performance (noise, input impedance, offset, etc...). <<<
Can you please explain why it has poor noise, poor inpt impedance, poor offset - and why these, other than noise, and to a very small degree, offset, are important in a MC phonostage?
>>I also miss the purpose and advantage of cascoding bipolars with tubes.
Was it designed following "subjective sound" criteria only? That would explain some of the decisions...<<<
I'm happyto amswer your questions after I'm clear as toyour complaints.
Regards, Allen
Syn, I believe you are focussing on engineering princlpals without regard to anything else, especially practicallitiy and sonic results.
We all have our preferences, you say "Transformers -yuk" and I agree with you. But you suggest a HV MOSFET and i say YUK. You use huge amounts of NFB to get distortions way below anything a human can sense - i personally HATE using NFB and have developed my circuits over 30 years without it.
The 22R input Rs are there to stop the whole mess oscillating at 30MHz or so.
The cascode bipolars/tubes were a R&D development when atube/tube casode was way too noisy for MC, and jfet/tube didn't have the 72dB gain in one stage (before RIAA eq) I needed. The MAT02 does the job and is VERY quiet in practice, Maybe not as quiet a 10 of them in parallel but more than enough for all cartridges I own, including a 100uV IKEDA.
The people who own this preamp love it, and most have owned serious commercial units before. It's going up against Joachim's best, how about sending one of yours over and there can be a real shootout?
I'm game!
Regards, Allen
We all have our preferences, you say "Transformers -yuk" and I agree with you. But you suggest a HV MOSFET and i say YUK. You use huge amounts of NFB to get distortions way below anything a human can sense - i personally HATE using NFB and have developed my circuits over 30 years without it.
The 22R input Rs are there to stop the whole mess oscillating at 30MHz or so.
The cascode bipolars/tubes were a R&D development when atube/tube casode was way too noisy for MC, and jfet/tube didn't have the 72dB gain in one stage (before RIAA eq) I needed. The MAT02 does the job and is VERY quiet in practice, Maybe not as quiet a 10 of them in parallel but more than enough for all cartridges I own, including a 100uV IKEDA.
The people who own this preamp love it, and most have owned serious commercial units before. It's going up against Joachim's best, how about sending one of yours over and there can be a real shootout?
I'm game!
Regards, Allen
Hello all !
I do not know where to start. Well, let´s say i learned it the hard way building the Hiraga Optime that a circuit with a serious technical flaw can sound excellent. It is not that bad if you look more closely. It´s low noise in my system and it is wide band. The input impedance works with my 5.5 Ohm Titan i without audible treble loss. It has quite high distortion. 0.32 % @ 1V RMS after the RIAA but ONLY 2nd montonic. It does not sound distorted to my and other trained ears. Quite the opposite. It sounds clean and dynamic. If you apply feedback you always end up with lower but more complex distortion. If i trust Geddes who spend his half life searching for correlation between subjective sound and distortion measurements the human ear is very tolerant to 2nd but higher order harmonics should be weighted much more aggressively. A simple THD reading says nothing so listening is required. That does not say that a circuit sounds poor with feedback. My Goldstandard Pre-Pre uses feedback but not the NFB parallel kind and has distortion below the -145dB noisefloor. In speed and precision it beats the Hiraga and needs so sorting, tuning or matching. High Z and low Z versions can be made. I would only condemn a product if it failed in my listening room. I will have a chance to listen to the Vacuum State and i am very interested to hear the latest Synaesthesia too.
It is technically brilliant and why should it sound bad ? There is still so much to learn until we reach perfection.
I do not know where to start. Well, let´s say i learned it the hard way building the Hiraga Optime that a circuit with a serious technical flaw can sound excellent. It is not that bad if you look more closely. It´s low noise in my system and it is wide band. The input impedance works with my 5.5 Ohm Titan i without audible treble loss. It has quite high distortion. 0.32 % @ 1V RMS after the RIAA but ONLY 2nd montonic. It does not sound distorted to my and other trained ears. Quite the opposite. It sounds clean and dynamic. If you apply feedback you always end up with lower but more complex distortion. If i trust Geddes who spend his half life searching for correlation between subjective sound and distortion measurements the human ear is very tolerant to 2nd but higher order harmonics should be weighted much more aggressively. A simple THD reading says nothing so listening is required. That does not say that a circuit sounds poor with feedback. My Goldstandard Pre-Pre uses feedback but not the NFB parallel kind and has distortion below the -145dB noisefloor. In speed and precision it beats the Hiraga and needs so sorting, tuning or matching. High Z and low Z versions can be made. I would only condemn a product if it failed in my listening room. I will have a chance to listen to the Vacuum State and i am very interested to hear the latest Synaesthesia too.
It is technically brilliant and why should it sound bad ? There is still so much to learn until we reach perfection.
Today i build another PSK buffer this time with LSK389C. The C version has higher IDss and that brought a well come gain in bandwidth (from 3.5MHz -3dB to 6.5MHz -3dB) and more class a. This time i am using the Alps K faders from studio consoles. I used 2 x 10kOhm log and put them in parallel. Balance can be adjusted easy that way. i already told you that i stole that trick from a Curl pre. Frequency response is now 2Mhz -3dB in the worst position of the faders, this time at around 3/4 full. The sound is now very transparent and open in the treble. The slightly dark character is gone. It sounds close to my reference with the exception of slam and extension in the bass. I attribute that to my primitive power supply. Just 4 9V alcaline cells plus 4 x 10.000uF Nichcon bypassed with 1uF MKC. It has a bit more DC offset, 6mV in both channels. No problem in my system. I use a safety cap in front of my power amp. This is really a great fun project, easy and cheep to build. Great sound and measurements.