I now use inverse parallel (anti-parallel?) as part of the Signal Ground to Audio Ground Disconnecting Network and as part of the Audio Ground to Chassis Disconnecting Network. This has to do with not being able to predict the Fault Current that might flow and thus what damage could be caused because excessive voltages build up across fixed impedances.
Hi all
The amplifire is working flawlessly now.
I am full of admiration over the elegant and perfectly working topology. In my veiw, The most elegant and well excecutet amplifier I have seen - DIY or otherwise.
I am in the process of evaluating the sound, but came to wonder if it is appropriate to post my thoughts in this forum?
There dosent seem to be many if any at all sonically describtion here though I guess a lot have been building these amps?
It will of course be my personal and subjective (how could it be otherwise) impressions .
What is your toughts concerning this?
Koldby
The amplifire is working flawlessly now.
I am full of admiration over the elegant and perfectly working topology. In my veiw, The most elegant and well excecutet amplifier I have seen - DIY or otherwise.
I am in the process of evaluating the sound, but came to wonder if it is appropriate to post my thoughts in this forum?
There dosent seem to be many if any at all sonically describtion here though I guess a lot have been building these amps?
It will of course be my personal and subjective (how could it be otherwise) impressions .
What is your toughts concerning this?
Koldby
After a few mental blocks on parts procurement I am making some progress on this amp. I presume that Linear systems LSK170C is a substitute for 2SK170BL, not that I have ordered any yet.
I found my digital multimeter cannot read the HFE of the 2SA and 2SC devices so I will have to resort to other means. Had a look at Rod Elliott's ideas on matching devices but I have to say I find them a bit half baked. Was wondering if I were to make (say) a 1mA current generator and use this to supply the base of the device under test in emitter follower mode. With Ib well defined we then measure the emitter current directly or calculate it from the voltage drop over the emitter resistor. Thus if we read 300mA the device has a gain of 300. Any comments?
Keith
I found my digital multimeter cannot read the HFE of the 2SA and 2SC devices so I will have to resort to other means. Had a look at Rod Elliott's ideas on matching devices but I have to say I find them a bit half baked. Was wondering if I were to make (say) a 1mA current generator and use this to supply the base of the device under test in emitter follower mode. With Ib well defined we then measure the emitter current directly or calculate it from the voltage drop over the emitter resistor. Thus if we read 300mA the device has a gain of 300. Any comments?
Keith
lsk170b = 2sk170bl
It is far better to use constant (or known) Ic when measuring hFE (Ib).
It helps keep each DUT at the same temperature.
The downside is that you don't get a direct hFE output.
I use a preset able CCS to fix the Ic and from the high side a high value VR + fixed R to measure the Ib.
adjust the VR until the whole batch of the devices can be measured without further adjustment.
Use SIL sockets or half a DIP socket to plug in the DUTs.
It is far better to use constant (or known) Ic when measuring hFE (Ib).
It helps keep each DUT at the same temperature.
The downside is that you don't get a direct hFE output.
I use a preset able CCS to fix the Ic and from the high side a high value VR + fixed R to measure the Ib.
adjust the VR until the whole batch of the devices can be measured without further adjustment.
Use SIL sockets or half a DIP socket to plug in the DUTs.
My initial impression of the FC-100 sonically:
It is a very VERY good sounding amplifier...!
I have on hand now for comparison two other amps:
A 6c33 based Output Transformer Less amp an a MOSfet amplifier witch is a further development of the Bow Technologies ZZ-8.
Tonally they are not that far from each other, the FC-100 the coolest of them.
Dynamically FC-100 is in the lead, closely followed by the 6c33. MOSfet is a bit slow.
The MOSfet amp is in this contex soundig a bit vieled and unprecise and is a little out of the contest.
A couple of examples of the difference between FC-100 and 6c33:
&c33 puts up a sligthly bigger room but the FC-100 differentiate the direct sound and the room answer better. (Beethove 7, Reiner, RCA living Stereo)
In the same album the cello/doublebass sounds more like they are made of wood in the 6C33, but clearer differentiation of cellos and doublebasses in the FC-100.
In choir music the S- sounds are better integrated with the voices in &c33, but differentiation between instruments / choir / room is better with FC-100.
(Cherubini Requiem, Kalmar Berkes, Hungarton)
(Stravinsky Petruska, Pierre Monteux, RCA Living Stereo) The 6C33 puts up a warmer sound and the flute sounds authentic on the other hand a little summary presentation compared to the FC-100´s ecxcelent handeling of this comlpex music.
(Rimskey Korsakov Scherasade Reiner RCA Living Stereo)
&c33 gets a little messed up and is a little less dynamic though the chymbals and triangle is full bodied . The FC-100 is dramatical and handels everything with ease and no messing up. The chymbals/triangle a fraction thin.
(Mahler 2 , Jarvi Virgin) The FC-100 can sound a little dry on the opening doublebasses compared with the warm woody sound of the 6c33.
(Mozart Pianoconcert 21 Reference Recordings)
In this modern, audiophile warm and spacious sounding CD, the FC-100 is bettering the 6c33 as the CD benefits from the very clear undistorted and pinpointing presentation.
(Tchaikovsky 6 Pletner DG)
6c33 presents a better sensation of the room while the FC-100 forshortens it a little and is sounding a tiny bit matter of fact.
(Siri Svales Jazzband Don´t Explain)
The vocalist's s-sounds is only handed well by the FC-100 (strange).
High Percussion is a little more fleshed out on the 6c33 nad the doublebass is a little warmer.
CONCLUSION:
(And again I must stress this is my personal veiw)
First: the 6c33 is NOT the old tube amp school: no sluppy bass and rolled of highs and chocolate midrange. It is the equal or better than all the high end maps I have heard!
Overall I think I would count the FC-100 as the best of the two, but some of the charateristiks of the 6c33 I would miss.
When I listen to the FC-100 I miss some of the warmth and mellowness of the 6c33 and when I listen to the 6c33 I miss the details the micro/macro dynamic and sensation of tistning to a presentation without distortion(!) or noise.
Hope you can use these impressions to something.
Comments wellcome.
Koldby
It is a very VERY good sounding amplifier...!
I have on hand now for comparison two other amps:
A 6c33 based Output Transformer Less amp an a MOSfet amplifier witch is a further development of the Bow Technologies ZZ-8.
Tonally they are not that far from each other, the FC-100 the coolest of them.
Dynamically FC-100 is in the lead, closely followed by the 6c33. MOSfet is a bit slow.
The MOSfet amp is in this contex soundig a bit vieled and unprecise and is a little out of the contest.
A couple of examples of the difference between FC-100 and 6c33:
&c33 puts up a sligthly bigger room but the FC-100 differentiate the direct sound and the room answer better. (Beethove 7, Reiner, RCA living Stereo)
In the same album the cello/doublebass sounds more like they are made of wood in the 6C33, but clearer differentiation of cellos and doublebasses in the FC-100.
In choir music the S- sounds are better integrated with the voices in &c33, but differentiation between instruments / choir / room is better with FC-100.
(Cherubini Requiem, Kalmar Berkes, Hungarton)
(Stravinsky Petruska, Pierre Monteux, RCA Living Stereo) The 6C33 puts up a warmer sound and the flute sounds authentic on the other hand a little summary presentation compared to the FC-100´s ecxcelent handeling of this comlpex music.
(Rimskey Korsakov Scherasade Reiner RCA Living Stereo)
&c33 gets a little messed up and is a little less dynamic though the chymbals and triangle is full bodied . The FC-100 is dramatical and handels everything with ease and no messing up. The chymbals/triangle a fraction thin.
(Mahler 2 , Jarvi Virgin) The FC-100 can sound a little dry on the opening doublebasses compared with the warm woody sound of the 6c33.
(Mozart Pianoconcert 21 Reference Recordings)
In this modern, audiophile warm and spacious sounding CD, the FC-100 is bettering the 6c33 as the CD benefits from the very clear undistorted and pinpointing presentation.
(Tchaikovsky 6 Pletner DG)
6c33 presents a better sensation of the room while the FC-100 forshortens it a little and is sounding a tiny bit matter of fact.
(Siri Svales Jazzband Don´t Explain)
The vocalist's s-sounds is only handed well by the FC-100 (strange).
High Percussion is a little more fleshed out on the 6c33 nad the doublebass is a little warmer.
CONCLUSION:
(And again I must stress this is my personal veiw)
First: the 6c33 is NOT the old tube amp school: no sluppy bass and rolled of highs and chocolate midrange. It is the equal or better than all the high end maps I have heard!
Overall I think I would count the FC-100 as the best of the two, but some of the charateristiks of the 6c33 I would miss.
When I listen to the FC-100 I miss some of the warmth and mellowness of the 6c33 and when I listen to the 6c33 I miss the details the micro/macro dynamic and sensation of tistning to a presentation without distortion(!) or noise.
Hope you can use these impressions to something.
Comments wellcome.
Koldby
AndrewT, thanks for your advice, particularly the bit about using part of an SIL or DIL socket to hold the DUT. The thicker leads on medium power transistors do not do the socket on the multimeter a lot of good. Anyone had experience selecting matched pairs of LSK170B as I am wondering how many I should buy to find a pair meeting Roender's specs?
Keith
Keith
I bought 600 of LSK170.
They all appeared to be in specification other than a few having Idss higher than the upper limit for grade C. I referred to them as C+ in swap meet.
I have used the a & b grades and found them to perform well.
I have sold a few hundred and so far no complaints on performance. I did supply a matched pair for a DCB1 that was reported in the Forum having an output offset mismatch that surprised me. I did not receive a complaint from the builder, probably because the offset difference was just a couple of mVdc.
Typical of FETs there is a wide spread of parameters when trying to measure them using my amateur methods. That spread makes matching of jFETs very much a try and see and hope for a sufficient yield. 100 samples gives a big margin for obtaining a good yield. This yield of matched pairs will depend very much on the tolerance required and the range of currents that the match still holds within that tolerance.
I would not expect to find any matched pairs in a batch of 12. You might get one pair of well matched and a couple of less well matched pairs in a batch of 20.
If you are only selecting Idss then expect a 98% yield. This effectively matches devices only when current is 100% of Idss.
They all appeared to be in specification other than a few having Idss higher than the upper limit for grade C. I referred to them as C+ in swap meet.
I have used the a & b grades and found them to perform well.
I have sold a few hundred and so far no complaints on performance. I did supply a matched pair for a DCB1 that was reported in the Forum having an output offset mismatch that surprised me. I did not receive a complaint from the builder, probably because the offset difference was just a couple of mVdc.
Typical of FETs there is a wide spread of parameters when trying to measure them using my amateur methods. That spread makes matching of jFETs very much a try and see and hope for a sufficient yield. 100 samples gives a big margin for obtaining a good yield. This yield of matched pairs will depend very much on the tolerance required and the range of currents that the match still holds within that tolerance.
I would not expect to find any matched pairs in a batch of 12. You might get one pair of well matched and a couple of less well matched pairs in a batch of 20.
If you are only selecting Idss then expect a 98% yield. This effectively matches devices only when current is 100% of Idss.
Last edited:
My friend METAL and I have chosen to build Miha's (Roender's) RMI FC-100 as our next DIY-project.
METAL and I have already successfully finished several projects (Carlos Mergulhao's DX Blame AMPs, different styles of Michael Bittner's SYMASM)
in this forum and have offered the results as group-buys.
Miha's RMI FC-100 is a new challenge.
METAL and I will work on Miha's EAGLE files (the AMP- and PSU files), we won't change anything on his layout (unless you tell us),
but make his tracks "smooth and elegant", so that you identify our EAGLE LAYOUT-handwriting.
This is the qualitiy that you can expect:
http://www.abload.de/img/to-3_symasym1hkhq.jpg
A double-sided etched, 2mm thick FR4 PCB, plated-through, the tracks being covered with electroless nickle/immersion gold, silkscreen on the component side, ...: the best quality that you have ever held in your hands.
If you are interested to build this AMP:
METAL and I will offer you in a group-buy:
- the AMP PCB (it will cost approximately 10€ per PCB, if we reach the quantity of 25 PCBs)
- the PSU PCB (it will cost about 7€ per PCB)
- I will try and get the NJL - transistors and offer them to you as well
But before we start, METAL and I need to know:
Are you interested to build this exceptional AMP?
To whose of you, who have already built the RMI FC-100: will you please give us your feedback about this AMP?
Is there anything that can be optimized?
I am waiting for your feedback.
If there is some interest, I will open a group-buy thread for the RMI FC-100.
Best regards - Rudi_Ratlos
METAL and I have already successfully finished several projects (Carlos Mergulhao's DX Blame AMPs, different styles of Michael Bittner's SYMASM)
in this forum and have offered the results as group-buys.
Miha's RMI FC-100 is a new challenge.
METAL and I will work on Miha's EAGLE files (the AMP- and PSU files), we won't change anything on his layout (unless you tell us),
but make his tracks "smooth and elegant", so that you identify our EAGLE LAYOUT-handwriting.
This is the qualitiy that you can expect:
http://www.abload.de/img/to-3_symasym1hkhq.jpg
A double-sided etched, 2mm thick FR4 PCB, plated-through, the tracks being covered with electroless nickle/immersion gold, silkscreen on the component side, ...: the best quality that you have ever held in your hands.
If you are interested to build this AMP:
METAL and I will offer you in a group-buy:
- the AMP PCB (it will cost approximately 10€ per PCB, if we reach the quantity of 25 PCBs)
- the PSU PCB (it will cost about 7€ per PCB)
- I will try and get the NJL - transistors and offer them to you as well
But before we start, METAL and I need to know:
Are you interested to build this exceptional AMP?
To whose of you, who have already built the RMI FC-100: will you please give us your feedback about this AMP?
Is there anything that can be optimized?
I am waiting for your feedback.
If there is some interest, I will open a group-buy thread for the RMI FC-100.
Best regards - Rudi_Ratlos
Last edited:
Great idea and initiative Rudi!
Provided the NJL power-transistors will be obtainable, I`m definitely very interested (in about 10 amp boards, depending of the PCB dimensions, maybe even more).
BTW: with "PSU-PCB", do You mean the (regulated) front-end supply?
While the silk screen component designations would be a big step forward the present layout does not have much spare real estate to allow this to happen. In fact the layout is so tight that there is practically no scope for parts substitution whatever.
Something I found using Panasonic 1,000uF capacitors was that there is not enough space to fit the output Zobel parts. Maybe the thru plated holes for the output devices need to be moved a few mm further apart. Either that, or move the Zobel parts along the board towards the main reservoir capacitors to avoid the squeeze between the pairs of 1,000 uF.
Keith
Something I found using Panasonic 1,000uF capacitors was that there is not enough space to fit the output Zobel parts. Maybe the thru plated holes for the output devices need to be moved a few mm further apart. Either that, or move the Zobel parts along the board towards the main reservoir capacitors to avoid the squeeze between the pairs of 1,000 uF.
Keith
It's time for a spreadsheet then:
https://docs.google.com/spreadsheet/ccc?key=0AjoWwC_y8PtfdG9OZEN0T25aeTlZcHpkcHhxVUQ2clE
Edit: See the following post.
https://docs.google.com/spreadsheet/ccc?key=0AjoWwC_y8PtfdG9OZEN0T25aeTlZcHpkcHhxVUQ2clE
Edit: See the following post.
Last edited: