| x-pro |
Hi all,
here is the link to the article (unfortunately, in Russian) I've published in 1999 in "RadioHobby" in Kiev about the original idea behing Creek Audio OBH-8 preamplifier. Since that time newer versions of this circuit are used in the current line of OBH preamps. However, this circuit is a good example of what could be done with few transistors. The sound of this circuit is good and it is very DIY friendly :) .
http://vlab.netsys.ru/audio/lpcor7.stm
It is single-ended class A with passive EQ. What else would you wish? :)
Technical Data of this preamp:
Gain - 90
SNR - better than 70 dB A-wtd ref 5 mV 1kHz input
RIAA deviation +/- 0.5 dB 50Hz-30kHz, -3 dB at 20 Hz
Max. output 5V RMS
Distortion for 1V output - less than 0.01% at 1 kHz
To prevent possible questions - it is my own design and the original publication was made with permission from Creek. The circuit is copyright of Creek Audio Limited (it is printed in the original article, unfortunately on the linked web site they cut the picture and lost it... .
Enjoy
Al |
|
|
| peranders |
| Gain = 90 at 1 kHz? |
|
|
| x-pro |
Yes, 90 at 1 kHz.
Al |
|
|
| Werner |
Per-Anders probably misread it as
90dB
Thanks for sharing the circuit with us. It is interesting, although for sheer simplicity nothing beats a Pacific.
 |
|
|
| hsort |
Hi X-Pro,
you mean about 40 dB.as input 3,5mV and output 300mV |
|
|
| x-pro |
2 Werner - yes, Pacific is simpler, however it does not have much gain and it's linearity is somewhat suspect :) . My circuit does have high enough gain and reasonably low distortion - even at 4-5 V RMS output, just before clipping, distortion are about 0.1-0.2% at 1 kHz. And this is using only one gain transistor Q1 - the rest are followers and a current source :).
2 hsort - it is 90, or about 39 dB at 1 kHz.
Al |
|
|
| Kuei Yang Wang |
Koinichiwa,
| quote: | Originally posted by x-pro
2 Werner - yes, Pacific is simpler, however it does not have much gain and it's linearity is somewhat suspect :) .
|
Linearity is arguable, though the harmonic structure of the Pacific is exemplary.
The gain is actually higher than 90, assuming the correct FET's are used. The 2SK170 @ 5mA has around 30mS (30mA/V) transconductance and operates into around 2k4 Loads. This gives a stage gain of 30mA*2.4k = 72. With 2 stages the overall gain becomes 5200 or 74db, of which a bit over 20db are lost in the passive RIAA EQ.
So the Pacific after accounting for the RIAA has around 60db Gain and around 0.15uV Ein (150nV Input Referred Noise) which makes it (marginally) suitable for medium output MC's without further Stepup or Pre-Pre. Output Impedance is around 2k. Not bad for a brutally simple circuit.
A friends Pacific runs fine and sounds VERY good. Originally it used 9V Block Batteries, when it stayed with me for a while (it drove a S&B TX-102 ever so much better than my Shindo copy Preamp) I converted it to using a 24V non looped switching supply, with better sound than we got from the batteries and no regular recharging!!!!
| quote: | Originally posted by x-pro
My circuit does have high enough gain and reasonably low distortion - even at 4-5 V RMS output, just before clipping, distortion are about 0.1-0.2% at 1 kHz. And this is using only one gain transistor Q1 - the rest are followers and a current source :).
[/B] |
Mebbe, but what does the harmonic structure of that circuit look like and what does it sound like? The original Creek phonostages sounded uninspiring.
Sayonara |
|
|
| Illusus |
I just finished building an "el Cheapo" for a friend (thank you for the advice). Even though I had to build it on a budget and skimp on a few components, it sounds fanastic, shockingly so.
I am in the middle of building a new turntable for myself. I want to build an intergrated pre, hardwired to the arm and interconnect. I chose the Pacific, because of it's simplicity it will allow me to keep the footprint small enough to fit into my arm base, plus I already have a boatload of the correct fets. I was wondering if you had any suggestions for the components to use(brands, models) in it's construction, cost is not much of an issue, but availability is, I hate hunting for parts and ordering from more than one or two sources.
Domo aregato,
Bart G. |
|
|
| Kuei Yang Wang |
Koinichiwa,
| quote: | Originally posted by Illusus
I was wondering if you had any suggestions for the components to use(brands, models) in it's construction, cost is not much of an issue, but availability is, I hate hunting for parts and ordering from more than one or two sources.
|
We build the whole circuit hard wired, in a copprfoil clad wooden enclosure. We used Holco Resistors (sorry, no longer available as non-magnetic, current Holco's are NDFG), LCR 0.47uF Foil & Film Coupling Capacitors (plus MIT 2u2 PPMF "bypass" on the output) and LCR Polystyrene Cap's for the EQ. This Parts Mixture is true "UK Style" and most components are/where available from Radio Spares and/or Farnell.
Sayonara |
|
|
| Cobra2 |
Have anyone tried this? comments?
I think it is from a Hafler Pre-Amp....
Arne K |
|
|
| Onvinyl |
Hi all,
just finished my version of the pacific (some pics are here I guess you will need an yahoo-account, for i have no own webspace)
It has an additional MC-Pre Stage. Thorsten is right, it has lots of gain (acutally a bit too much)
As a friend stated, it sounds that 'you miss nothing'.
(Same night we tried the blue amp phonostage, well, there is still a lot to improve...)
It is a good value for money, but, on the other hand, no 'beats everything, costs nothing'. Of course, other implementations may sound different.
Rudiger |
|
|
| Kuei Yang Wang |
Koinichiwa,
| quote: | Originally posted by Onvinyl
(Same night we tried the blue amp phonostage, well, there is still a lot to improve...)
|
Nice work in general, but the ouput coupling cap is a little prosaic.
You can reduce the Drain Load resistors of all stages to reduce gain. This will also allow you to "tune" the sound somewhat. Just as with Pentodes, the J-Fet is more linear into lower loads.
However, the biggie (if you ask me) is to add a third battery and to cascode the two J-Fets with a nice bipolar transistor each, base connection directly to 12V, +V 36V. This makes the drain load of the J-Fet almost zero, maximising the linearity of the circuit.
You may be able then to increase the Gain further by adding the +V up to 48V and doubling or even trippling the load resistor, which may allow you to remove the MC Pre-Pre stage.
Sayonara |
|
|
| Onvinyl |
Hi,| quote: | | ... , but the ouput coupling cap is a little prosaic. |
Quiet true
:goodbad:
| quote: | | You can reduce the Drain Load resistors of all stages to reduce gain. |
At least the drainres of the stage before the riaa is part of the network, if I understood it right, so tuning might outtune things...
| quote: | | However, the biggie (if you ask me) is to add a third battery and to cascode the two J-Fets with a nice bipolar transistor each, base connection directly to 12V, +V 36V. |
That may be worth a try.
thanks,
Rudiger |
|
|
| Marinos |
Hi all,
Some months ago I have built a copy of Pacific and I really like it. The gain is more than enough for my high output VDH MC2.
I used parts from my stock ( beyschlag mf 0.6w resistors, ICL MKP capacitors, SCR tin foil 100n, 2 SLA 1.2Ah // a Mallory 12000uF ) , and I must tell you that this little thing really sings..
Later I added a 470uF decoupling capacitor + 10R carbon film isolation resistors per fet , 100R carbon film "gate stoppers" and finally stacked a 110R resistor on the 30n RIAA capacitor (hidden tc).
All of these changes improved the sound especially the decoupling capacitors.
Now I'm finishing a second build using better parts ( Maplin M mf resistors, SCR tin foin caps, silver mica 30n, Elna Starget 470uF )and better layout but I have two questions.
- The prototype plays ok with the 100K input resistor but the recommended loading for my cartirdge is 47K. Should I change it to 47K or I'm altering anything important for the operation of the 1st stage ?
- The correct position of the resistor for the hidden tc is on top or at the bottom of the 30n cap ?
Any other suggestion/advice is welcomed.
I'm learning but slowly......
Thanks in advance for your help
Marinos |
|
|
| Onvinyl |
Hi,
never heard about adding a resistor for time constant.
But another thing: I recently wrote, the pacific didn't reach the awaited sonic level. Meanwhile I replaced the accus with big panasonics (7,2Ah each) -- it is a completly different thing now.
Sound is deep, tight and sweet at the same time. Not bad, really!
Rudiger |
|
|
| Jocko Homo |
Yuk. Built one once for someone. The RIAA accuracy was poor. Some may like its "harmonic structure", but I didn't. A better solution would put a follower in between stages.
Jocko |
|
|
| Marinos |
Hi Rudiger,
Adding a resistor to the HF part of the filter you stop boosting the HFs at around 50KHz (3.18 ìS). This has been referred as the 4th or hidden time constant. See this link for more information.
http://www.klaus-boening.de/html/ti...l#time constant
Experienced guys, like Werner in this forum, have mentioned the difficulty of making a suitable PSU for pacific. I started with batteries and from the beginning I have really good sound from it.
Using 7,2Ah SLAs you start with a clean supply, with an internal resistance around 20 mOhm, and capable of huge transients.
I think that this is what you hear now.... Do you bypass the batteries with a BIIIG capacitor ?
Sorry but what do you mean with 'accus' ?
Marinos |
|
|
| Marinos |
Hi Jocko,
I also think that something is wrong with RIAA values. With 27K, 3K, 100n and 30n the sound is bass heavy and a little splashed highs. I checked the values with TCJ's RIAA calculator and I came to the following better sounding values 24K9, 3k2, 100n, 35n.
These values are very close to Werner's Pacific variation (see attachment). What values did you use?
For the Pacific "as is" I will only check if moving the 470n coupling capacitor after RIAA sounds better.
For a "modified" Pacific I would like to implement Thorsten's suggestion (see this post, BJT cascode....). Sorry but I don't have the background... to understand the " ...a follower in between..". Could you please explain this in more detail ?
Any better RIAA pre suggestions ?
Marinos |
|
|
| Onvinyl |
Hi Marinos,
with accus I meant batteries, and yes, these are bypassed by huge amounts of uF...
from my limited experience I might add the following:
1) If you use 1.8K Drain-Resistors instead of the mentioned 2.4 K, you change the riaa network as well! Lowering the 27K might cure this. You will as well get lower gain with 1.8
2) I really made a big effort to decouple the very first stage (head for MC-pickup) (choke, Res, couple of caps). That paid!
3) what's the point of the 2.2megs?
Suggestions for better? Currently, I'm building la bohème, we'll see...
Thorsten's suggestions seem to be inspired of klaus boening's design, I think...
Rüdiger |
|
|
| Kuei Yang Wang |
Konnichiwa,
| quote: | Originally posted by Onvinyl
Thorsten's suggestions seem to be inspired of klaus boening's design, I think... |
I think not.
We added cascoding Transistors to the Pacific long before Mr. Boening published his design.... I have liked using Cascodes in Solid State Circuits for a long time as they allow to combine devices to achieve normally conflicting performance items.
The simplest way to add the cascoding is to use good (Low Noise High Ft, High Beta, good Beta Linearity) NPN Transistors like BC239 and use a stack of 3pcs 12V Batteries for the Supply. The BJT's are added between FET Drain and Drain resistor with the Emitter to the Drain and the Collector to the Drain resistor.
Simply connect the Base of the Transistor(s) to the top of the first Battery (+12V) and apply +36V (nominal - in reality over 40V) to the +B Line. Worked a treat for us. BTW, the Drain Resistor should drop around 12V and the 2SK170 MUST be selected for identical currents for the two channels plus should have Idss of 4 - 6mA....
If one wants to implement Jocko's suggestion to place buffers simply add another BC239 Transistor with another BC239 as CCS. The Base of the first BJT goes to the signal takeoff Point, Collector to +B and the signal is taken of the Emitter. The RIAA needs re-calculating then.
I have done this and attached what I would suggest in this case.... Not as simple as before, but still a minimal number of components per gainstage (6components) and all the transistors can be glued together and hardwired so that hardwiring the circuit in air is still a perfect possibility. And each such gainblock offers a very low noise level with a gain of around 36db and pretty decent (thanks to cascode) linearity.
Another Idea, the schematic as posted here would be usable with the S&B 600R RIAA Module.
Simply connect a 600R resistor between the output of the first stage and the RIAA Module Input and remove all the other RIAA components. As the RIAA is input terminated the RC combo coupling to the next stage can remain just as shown....
Sayonara |
|
|
| analog_sa |
| More variations - unknown japanese origin |
|
|
| Onvinyl |
| quote: | Konnichiwa,
quote:
Originally posted by Onvinyl
Thorsten's suggestions seem to be inspired of klaus boening's design, I think...
I think not.
|
hu, no offence meant. In what respects does the new circuit sound different?
Rüdiger |
|
|
| Kuei Yang Wang |
Konnichiwa,
| quote: | Originally posted by Onvinyl
hu, no offence meant. In what respects does the new circuit sound different?
|
We have only build this cascoded, I simply drew in the Buffers as well and re-did the RIAA as it was discussed.... The cascoded circuit with the correct RIAA sounds cleaner than the original, more open and can handle complex passages better. I normally try to leave buffers out as much as possible as they rarely improve the sound.
Sayonara |
|
|
| analog_sa |
Kuei
What is your stance on splitting the riaa? If one is to have a third stage for extra gain would it be beneficial to separate the timeconstants? I seem to always prefer the sound when they're separated but usually there are way too many variables to make a direct comparison. |
|
|
| Kuei Yang Wang |
Konnichiwa,
| quote: | Originally posted by analog_sa
What is your stance on splitting the riaa? If one is to have a third stage for extra gain would it be beneficial to separate the timeconstants? I seem to always prefer the sound when they're separated but usually there are way too many variables to make a direct comparison. |
If you have three stages spliting the timeconstants can be better. Adding a stage only to split the RIAA is not a good idea however. Also, there are several ways to realise a "one go" RIAA and some sound better than others. I prefer the style applied by Kondo San in his Phonostages....
Sayonara |
|
|
| Marinos |
Hi all,
Thank you for your replies
Kuei,
another setup could be
- Simple cascode for the 1st stage and subtracting the drain resistor value from the first RIAA resistor.
- Cascode + follower for the second in order to have better drive capability for the interconnects and lower output impendance.
Right ?
Adding a selector, a pot and a 3rd Cascode + Follower we have a very nice and compact (psu in a separate case) Preamplifier.
Right ? |
|
|
| Kuei Yang Wang |
Konniciwa,
| quote: | Originally posted by Marinos
another setup could be
- Simple cascode for the 1st stage and subtracting the drain resistor value from the first RIAA resistor.
- Cascode + follower for the second in order to have better drive capability for the interconnects and lower output impendance.
Right ?
|
Right.
| quote: | Originally posted by Marinos
Adding a selector, a pot and a 3rd Cascode + Follower we have a very nice and compact (psu in a separate case) Preamplifier.
Right ?
|
You would need a different FET (IMHO) for the Linestage, lower transconductance and higher gate cutoff Voltage, a linestage with 36db gain re than a little excessive if you ask me....
Sayonara |
|
|
| Marinos |
Hi kuei
Have you tried something similar?
Any suggestions for such a FET? Are the europian OK or should I search the Toshiba family ?
Could such a Line stage compete sonicaly my SE 6BX7 preamp ? |
|
|
| Werner |
| A sketch that reflects my present feelings about this kind of circuit. Optimised for the 100nF MKVs I have. YMMV. |
|
|
| peranders |
| Q1, Que? I'll gues the schematic is drawn wrong? |
|
|
| Christer |
| quote: | Originally posted by peranders
Q1, Que? I'll gues the schematic is drawn wrong? |
No, it has a purpose. Think about it. :) Otherwise, you can find
it in an earlier thread, I just don't remember which one now. |
|
|
| peranders |
| What is Q1 doing then? |
|
|
| Christer |
The point is to try getting a constant supply current draw to
the input stage. It is just a dummy load intended to approximate
the inverse current draw of the input JFET.
I see now that Werner has edited and modified the schematic
slightly. It was perhaps a little bit more obvious in the earlier
one where V1 was a chain of RC links from the supply. |
|
|
| analog_sa |
| quote: | | It is just a dummy load intended to approximate |
Fascinating. Is the effect clearly audible? Why is C1 so small? |
|
|
| Christer |
| quote: | Originally posted by analog_sa
Fascinating. Is the effect clearly audible? Why is C1 so small? |
Well, don't ask me, I haven't tried it. Achieving a constant current
draw should be beneficial for low level amps in general though.
As I said, there was a thread a while back discussing this
general principle and various ways to achieve it. As I remember
Jonathan Carr was quite enthusiastic about it, saying that the
major advantage was that one can get away with simpler
power supplies. He said that the choice of caps in the PSU
was not so important anymore.
I don't know about the C1 value. Maybe a typo? Werner?
This is not only relevant in audio. We used to have a company
called Facit here in Sweden which was quite big and advanced
in computer peripherals. In the 70's they tried to get an order
for printers to Pentagon. However, Pentagon did not accept the
printers after testing them, since they claimed it might be possible
to monitor the noise on the mains line and figure out what was
printed. These were dot matrix printers, so Facit made a special
military version of it, by simply adding an extra dummy print
head which was fed with the complementary data of the ordinary
one. The power demand thus summed to a near constant one. |
|
|
| Marinos |
Hi all,
Werner explains this setup at his site
http://www.geocities.com/Hollywood/...133/design.html
"........Since this scheme would result in a high power supply source impedance at low frequencies, the supplies load must be kept constant relative to the music signal. This is done hanging a source follower off the first stage's transistor which runs at the same current, but inverse in polarity. The follower can be filtered at its gate to restrict its operation to low frequencies only. ".
I think that the last sentence explains C1 operation.
Don't forget to give me some help about the "integrated preamp" (see previous post) guys. |
|
|
| Werner |
No. The load at Q1's source is there to simply mirror the load at the front-end transistor's drain.
That 100nF and 750R at Q1 stems from part of the RIAA filter. We don't copy the whole RIAA as the bulk of the current flows in the 750/100nF pair anyway.
I make no claims as to the sonic benefits of this scheme. I expect it to depend heavily on the type of voltage supply anyhow. Some of you may want to try it, though. |
|
|
| peranders |
| quote: | Originally posted by Werner
I make no claims as to the sonic benefits of this scheme. I expect it to depend heavily on the type of voltage supply anyhow. Some of you may want to try it, though. |
Werner we have a discussion about the Jung regulator with extremely low output impedance (and the importance of it), your power supply seems to be very week since you need current compensation?? Isn't low impedance of the power source essential at all times? |
|
|
| Werner |
| quote: | Originally posted by peranders
your power supply seems to be very week since you need current compensation?? |
Strange, but I can't remember having disclosed anything about 'my power supply' in the above thread.
The schematic was an indication of a concept, and as such, no power supply is attached to it. |
|
|
| peranders |
OK, but suppose you HAVE a very powerful voltage source like a Jung, Sulzer regulator or even a big fat cap, why trying to draw constant current?
At work have a similar thing but in that case it's about discharging smooting caps because the current consumption is very unequal when a relay is on or off. |
|
|
| peranders |
| quote: | Originally posted by x-pro
Hi all,
here is the link to the article (unfortunately, in Russian) I've published in 1999 in "RadioHobby" in Kiev about the original idea behing Creek Audio OBH-8 preamplifier. Since that time newer versions of this circuit are used in the current line of OBH preamps. However, this circuit is a good example of what could be done with few transistors. The sound of this circuit is good and it is very DIY friendly :) .
http://vlab.netsys.ru/audio/lpcor7.stm
It is single-ended class A with passive EQ. What else would you wish? :)
Technical Data of this preamp:
Gain - 90
SNR - better than 70 dB A-wtd ref 5 mV 1kHz input
RIAA deviation +/- 0.5 dB 50Hz-30kHz, -3 dB at 20 Hz
Max. output 5V RMS
Distortion for 1V output - less than 0.01% at 1 kHz
To prevent possible questions - it is my own design and the original publication was made with permission from Creek. The circuit is copyright of Creek Audio Limited (it is printed in the original article, unfortunately on the linked web site they cut the picture and lost it... .
Enjoy
Al |
Anyone who knows where the schematic went to? The page above is removed, or moved. |
|
|
| uniuniunium |
| so is this a clone of the current production OBH-8? |
|
|
| x-pro |
| quote: | Originally posted by uniuniunium
so is this a clone of the current production OBH-8? |
It is the circuit of OBH-8 as it was in 1996 :) . Current production may be a bit different.
x-pro |
|
|
| pekka |
X-pro, thanks for the interesting schematic. The Creek phono preamp has consistently received quite favourable reviews in the magazines I have read.
Unfortunately the picture resolution makes is difficult to read some of the resistor values. I'd rather get them directly from the source than guess them myself...
Could you perhaps post the values of:
R2, R3, R7, R9, R11, R14 and R16.
The transistors are BC550B, I suppose?
Best regards,
Pekka U. |
|
|
| x-pro |
| quote: | Originally posted by pekka
X-pro, thanks for the interesting schematic. The Creek phono preamp has consistently received quite favourable reviews in the magazines I have read.
Unfortunately the picture resolution makes is difficult to read some of the resistor values. I'd rather get them directly from the source than guess them myself...
Could you perhaps post the values of:
R2, R3, R7, R9, R11, R14 and R16.
The transistors are BC550B, I suppose?
Best regards,
Pekka U. |
No problem:
R2 = 150K
R3 = K62
R7 = 68K
R9 = K33
R11 = 100K
R14 = 22K
R16 = K22
Transistors are BC550B or C . It may be nesessary to adjust the value of R5 (1M2) in the range from 1M0 to 2.2M depending on the actual make of transistors, otherwise there could be some deviation from correct RIAA curve below 50 Hz.
x-pro |
|
|
| pekka |
'preciate it!
Pekka |
|
|
| audionline |
| Hello,X-pro,Can you give me X-pro's preamp's schematics?I can't give it ,,Thank you very mach ... |
|
|
| audionline |
| MY E-mail:mc_wilson@21cn.com |
|
|
| audionline |
| Thank you very mach ..Good luck everyday... |
|
|
| arnoldc |
| Hi x-pro, did you design the OBH-15 too? Are they discrete like the 8 or opamp based? |
|
|
| x-pro |
I'll have a look, thanks.
| quote: | Originally posted by arnoldc
Hi x-pro, did you design the OBH-15 too? Are they discrete like the 8 or opamp based? |
No, OBH-15 was made after I've left and I have no information about what is inside, sorry.
Cheers
Alex |
|
|
| Wizard |
| Mr.Alex,Thank you very mach!!! |
|
|
|
