Bruce Heran   |
The Groovewatt Tube (Valve) RIAA Phono Preamp Project
This is a project that I always wanted to do, but managed to put off for a long time. It has been over a year in the doing. Partly because there are so many good phono preamps in existence and I was slightly concerned that it was beyond my design skills. It took over a year to get it up to my satisfaction. There were several iterations and lots of frustration along the way. My goal was to make a preamp that would not embarrass the builder with inferior performance. I used a number of commercial products for comparison. My thought was why bother to design and build something that was already available for a lower cost. As it turns out, it was not as inexpensive nor un-complicated as I first hoped. I wanted response within 0.5 dBV of the RIAA curve, low distortion, and a signal to noise of about -90 dBV. The first two were easy, the signal to noise was not. I did a lot of research on the RIAA curve and the various problems that are inherent in phono preamps. I tried active, passive and a combination of active and passive equalizations. All work and all have positive features and all have negative features. I tried several configurations and tube types. The final design uses things I am familiar with. The input stage is a symmetrical SRPP. It feeds the passive equalization network. The second stage is nearly identical with the first and is direct coupled to the third stage which is a cathode follower. I eventually used Gold Pin JJ ECC83S in both of the SRPP stages and half of a ECC81 (12AT7) for the cathode follower. Gain is sufficient at 45 dB for me to use a high output MC cartridge (I use a Dynavector 10X5 with an alternate of a Denon DL-110) to drive a passive volume control to feed a pair of Oddblocks. The initial design tried using other tubes but none had sufficient gain in the SRPP configuration. The performance of JJ ECC803S tubes is nearly as good as the Gold Pin JJ ECC83S and is a lower cost option. Be sure to ask for matched pairs though. The third stage is almost optional as the output impedance of the SRPP is low enough (14K) to drive most line stage preamps and directly to power amps with a 100K passive volume control. I added that stage so that it would be possible to drive long cables and following equipment that had low impedance inputs. It is actually possible to drive some high sensitivity 300 ohm impedance headphones to sufficient volume. The preamp is designed for moving magnet (MM) and high output moving coil (MC) cartridges with a signal level over 2.0 mV. I believe the best way to handle the output from a low output MC cartridge is with a transformer so no attempt was made to handle them with this project. On the schematics I show a way to get as much as 3 dB more gain, but at a slight expense of more noise. The name of the project came from another DIYer (Thanks, Shane). I asked for suggestions and his was the best. He won an "Oddwatt T-shirt" for the idea.
Photograph 1: Groovewatt Type III DIY Tube RIAA Phono Preamp
Warning: This project uses dangerous and potentially lethal voltage levels. If you are unfamiliar or do not feel comfortable working with the voltages involved, I highly recommend that you do not attempt the project.
Design Goals - Tube (Valve) RIAA Phono Preamp Project
The actual design is almost modular in that both gain stages are nearly identical. Circuit values in the equalization section are the exact ones needed and in the case of the resistors you will probably need to series a few smaller ones to get the needed values. Some deviation is possible and while it will alter the RIAA response, the actual amount of deviation will most likely be inaudible. I found 11k resistors at Parts Express
which in place of the 10.9k will not mess up compliance with the RIAA curve much. The capacitors should have tolerances as close as you can get. I used a capacitor meter to select mine from a large batch of samples. With both the resistors and capacitors, the closer to the values the better the performance will be.
I used all audio grade (in this case Jantzen) capacitors as bypass capacitors and only quality electolytics. The output capacitor should be a "poly" cap for best results. The ones I used were Jantzen Silver Z as I like the way they sound. These are physically large so keep that in mind when designing your layout. For the inter stage coupling capacitor I used an Auricap however in both cases, other quality brands ought to be satisfactory. In both instances the capacitor will tend to "flavor" the overall sound. If you know how they sound and like some other brand, by all means use them.
Parts can be obtained from sources like Parts Express, Jameco and MCM Electronics,
among others.
In an earlier build I wanted to have the capability to do mono records as well as stereo. So I included a switch that connects the two channels at the input (shown in the photo below of an earlier build).
Photograph 2: Groovewatt Valve RIAA Phono Preamplifier with protruding tubes
There are actually two different preamps shown in the photos. The one with the tubes on top of the chassis (Photograph 2) is the first one built and the preamp shown in Photograph 1 is the third build. The first build shows how to separate and shield the various preamplifier sections for low noise (Photographs 5 to 7). The third build incorporates shielding differently and has a more compact arrangement of the active circuitry. I can not overemphasize the importance of shielding and good grounding practices with this project. See my suggestions regarding grounding and shielding for your DIY audio projects. The two preamplifiers are electrically the same with the exception of a different value and brand of output capacitor. The original version used a 4.7uF WIMA and the later one a 1.0 uF Jantzen Silver Z for the output coupling. Unless you need to drive a low impedance (below about 10k) preamplifier the 1 uF output capacitor is quite sufficient to allow response down to the 20 Hz range. Another difference is the use of Gold Pin JJ ECC83S tubes in place of the JJ ECC803S. The JJ Gold Pin ECC83S tubes were about 1 to 1.5 dB quieter. This helped me reach the -90 dBV level of noise I sought.
Power Supply - Tube RIAA Phono Preamp
The power supply is fairly straight forward in design. It uses an inexpensive transformer from Edcor. A toroid would be most likely be better, but I could not find one that was inexpensive and suitably small. One I did find was nearly three pounds and about 5 inches in diameter. The high voltage is rectified by discrete solid state rectifiers. The specified ones (UF4007) are a fast recovery variety and standard rectifiers should not be used. The filter section is divided into two separate strings, one for each channel. This is done to enhance channel separation. Electrolytic filter caps are bypassed by audio grade polys. In one version I used LR8N solid state regulators in the high voltage string. They worked fine, but they did not seem to improve the overall S/N and the complication I felt was unjustified. Portions of the individual channel filters are outside the main power supply housing and can be seen on the rectangular PCB near the front of the chassis. An LED power indicator is located just under the right side of the top grill. The grill BTW came from a hardware store and is a section of "kick panel" for a screen door.
Figure 1: Power Supply Schematic - Groovewatt Tube (Valve) RIAA Phono Preamp
Heater Supply - Tube RIAA Phono Preamp
I employed DC on the heaters. The 8.5 volt winding on the transformer was rectified by a full wave bridge and filtered. Then it is routed through a single solid state rectifier (to drop the DC voltage by approximately 0.7 volts) and a 1.5 ohm resistor to a second filter and then applied to the tube heaters. The value of the resistor shown on the schematic is 0.5 ohms and the actual value can be in between those values. It is somewhat dependent on the rectifiers, filter capacitors and transformer. You could of course use a 6 volt regulator like the 7806. It would need to handle about 1.5 amps and require a heat sink. The heater circuit MUST NOT be grounded. I apply a positive voltage to the heater circuit of approximately 80 volts to protect the SRPP tubes from excessive cathode to heater potentials. They will fail if you don't do this. The positive voltage is obtained from a divider string originating from the B+. Two stages of filtering were necessary to reduce the noise level to satisfactory levels. Do not omit the poly caps in the divider string and it is unnecessary to use larger values. There is virtually no current drawn by the divider string (over the small amount used by the dividers themselves) and there is no return path from the heater circuit. The voltage only provides a reference for the heaters. The photo below shows the power supply. Note that the power supply is contained in an additional aluminum enclosue and the cover is removed in this photo. The PCB in the photo is from another project.
Photograph 3: Power Supply - Groovewatt Tube (Valve) RIAA Phono Preamp
Construction - Tube RIAA Phono Preamplifier
This is a high gain circuit and needs to be treated as such. Open loop gain (ignoring the loss of the RIAA section) is in the 60-70 dB range and thus it is quite sensitive to picking up noise and hum. Layout is important. The signal path from input jack to the first SRPP should be short. If more than about 1-2 inches (25 to 50 mm) it needs to be shielded. The components in the passive equalization section need to be separated from the power supply. I suggest that the chassis be partitioned off into sections. Power and signal wires and components should be separated when possible and cross at right angles when they can't be separated. The power transformer could be problematic as it is unshielded. The use of tube shields is essential even though the tubes are inside the chassis as they are prone to picking up ambient noise. The main components of the power supply are contained within a metal housing. Even the AC switch is inside that area. There is a single point where all the grounds on the chassis come together near the power entry. I always use three wire IEC connectors and a combination AC input/line filter and integral fuse assembly. You can choose to have a power switch or not. That would depend on your system. If you choose not to have one, a B+ standby switch would be in order. Mine has a mains switch.
Figure 2: Schematic - Groovewatt Tube (Valve) RIAA Phono Preamp
For additional tube amplifier design and construction tips, see the design and construction tips and suggestions for vacuum tube amplifiers. Also, I have posted some suggestions for a tube amplifier wiring color code that you can use during construction.
Photograph 4: Inside View with PS Cover in-place Groovewatt Tube RIAA Phono Preamp
Assembly Verification and Startup
First look for loose wires and components. Turn the chassis upside down and shake it. You might be surprised how much junk falls out. Insert the fuse and tubes. Do not connect to any stereo gear (just in case there is a catastrophic fault). Connect the power cord to the preamp and turn the switch on. Do not connect the cord to the AC mains until later. This insures that you will not become part of the circuit in the event there is a short to the chassis. Attach a voltmeter to pin 6 on the first stage SRPP. This is the mid point and is the best indication of proper initial operation. Set the range (if not an auto sense meter) to 300-500 volts. Plug the power cord into the AC socket. Watch the meter reading.
It should be in the range of 80-120 volts after about 30 seconds. If it remains zero or is greatly outside that range, remove the power cord from the socket and look for some sort of wiring error. If it is in the proper range move the "hot" probe to teach of the SRPP tubes (pin 6) in turn all should be in that general voltage range. The variation is because of gain differences between the two triodes in the tubes. Then check pin 8 and pin 3 on the ECC81. The reading should be also in the same range. If any readings are way off, or there is arcing, sparks, etc, disconnect the power cord and look for errors. Next check the voltage on the heater circuit. Between pins 4 or 5 and 9 on the ECC83s it should be close to 6 volts DC. A little (up to about 0.25 volts) above or below is fine. Then measure the DC voltage from pin 9 on any of the tubes to the B+ ground. It should be between 70 and 90 volts. Any time any of the readings is greatly different you should stop and determine the cause. If all readings fall into the proper ranges you can now try the preamp in your system. If there is excessive hum, noise or no sound, then you need to go back and find the error.
The photographs below show the inside views of the earlier Groovewatt build with protruding tubes shown in Photograph 2.
Photograph 5: Top View PS Cover Removed - Groovewatt Tube RIAA Phono Preamp
Photograph 6: Inside View Covers Removed - Groovewatt Tube RIAA Phono Preamp
Photograph 7: Inside View Covers In-place - Groovewatt Tube RIAA Phono Preamp
Measured Performance - Groovewatt Tube RIAA Phono Preamp
Testing this preamp has been a challenge. I tried many ways to verify that the RIAA standard was met and actually what was the S/N and such. I don't know that what I present below meets "industry standards" but it should give you some feel for the overall performance.
The performance testing equipment setup consisted of:
- HP 331A Distortion Analyzer
- Tenma 25MHz Dual Channel Digital Storage Oscilloscope
- Velleman Dual Channel USB PC Oscilloscope
- Tenma Low Distortion Signal Generator (0.05% residual THD)
- APC H10 Power Conditioner (120VAC output)
- Several digital volt meters
Distortion: (including hum and noise): 0.6% to 0.7% at any frequency between 20 Hz and 20 kHz at 1 volt RMS output into 10k resistive load. The distortion is unchanged up to over 20 volts of output.
Signal to Noise including hum: Better than 85 dB below one volt RMS output, wide band, unweighted. Measurements on the second build indicated a Signal to Noise Ratio (SNR) of approximately 91 dBV.
Frequency Response: Within 0.3 dB compliance with the RIAA standard from 20 Hz to 18 kHz. It was minus 0.8 dB at 20 kHz and minus 3 dB at 27 kHz. The deviation at the upper frequency extreme is somewhat tube dependent as different brands and types (e.g. NOS 12AX7s, versus new manufacture European spec ones) have somewhat different response curves in part due to variances in construction and in some cases miller capacitance.
Gain: At 1 kHz the gain is 45 dBV.
Impedance: Input impedance (and capacitance) is set by dip switches to match various cartridges. Output impedance is approximately 250 ohms.
Listening Notes - Groovewatt Tube Phono Preamp
As always your results may be different from mine and what and how you listen is likely to be different as well. My main system consists of a pair of Martin Logan Vistas (ESLs) driven by a pair of Oddblocks using Blue Glass JJ EL34L tubes, A pair of 7 cubic foot subwoofers with 15 inch drivers powered by an Amp 100 (so I use a piece of SS gear- for this application it works), signal separation is handled by a Marchand 24 dB/octave electronic crossover set for 50 Hz, a Forewatt tube line stage preamp (usually set in passive mode), an Oppo BDP83SE for a digital source, a Dual 701 direct drive turntable with a Dynavector 10X5 MC cartridge and for a back up phono preamp a Simaudio Moon LP3 (for comparison with the GrooveWatt). Audio interconnects are by Audioquest. With the GrooveWatt in place, the sound is exceptionally smooth, but not lacking in details. It is full range with solid well defined bass. Soundstage is approximately as wide as the Moon LP3. Only at full gain using the preamp in active mode is there any background noise. The music flows smoothly from a completely dark background. It is quieter than any LP I own. The Moon is quieter, but then it is among the quietest phono preamps available (at about -105 dBv).
Additional Information
As with all DIY projects, it is as much a process as an end result and I fully expect to make some refinements to it. So stay tuned to the Groovewatt tube RIAA phono preamp support and update thread in the Forum where the changes (if any) will show up first. You can also use the thread to ask questions or post comments. This is not a simple project, but equally it is not at the expert level. It lends itself to some tube rolling. I tried 6N1P tubes for the cathode follower and did not like the sound as much as the ECC81 types. Different brands will color the sound as well. Be sure to use low noise and matched gain tubes for best performance. I hope this project will give you as much pleasure building and listening as it does for me.
UPDATE - 22 November 2011
Bruce's shares detailed listening notes for 6 phono preamplifiers including the Groovewatt. See an Informal Comparison of Six RIAA Phono Preamplifiers for full details.
UPDATE - 11 January 2015
Since DIYers are a different breed from kit builders most of the time...especially when the kit will be after the assembled ones are on the street by about a year, I am posting the schematic for the one that is in the final design stage for commercial sale. The Groovewatt phono preamplifier kit is still a few months away, but the design is firm (at least as any of mine are). So here is the schematic. With a careful building it is top notch. Alternative step up trannies can be used, but the Lundahl LL1678 seem fine IMO. The build will require some understanding of how to build high performance tube gear. Parts placement, wiring methods and general layout are really quite important for success. If you do it right, it will run with the big boys. I don't recommend this project for novice builders.
Figure 3: Updated Schematic - Groovewatt Tube (Valve) RIAA Phono Preamp
Figure 4: Updated Power Supply Schematic - Groovewatt Tube (Valve) RIAA Phono Preamp
Two adjustments can be made. First while the preamp has nearly 45 dB gain you can get a little more by using quality 100 uF / 16 V capacitors on the cathodes of both gain stages. You will get an additional gain of approximately 5 dB. There will be a slight noise penalty, but it is still IMO acceptable (over 90 dBV S/N). The second possible change is the capacitor indicated with a red arrow on the schematic. The use of this part is dependent on how good your grounding system is and to some extent layout. It can be optional, attached as shown or attached to the other leg of the heater circuit. In DIY builds it is pretty much impossible to determine the correct attachment. In all my builds it is as shown. It has the potential to improve the S/N by about 1.5 to 2 dB. Quite worthwhile IMO.
If you need to use the preamp with a line stage that has an input impedance below 10 k-ohm I would increase the output coupling capacitor to about 5 uF to ensure proper low bass response. If you use different tubes the gain may differ as it is open loop (no global NFB) and possibly some slight shift in the very high end response due to their different inter electrode capacitance values. Noise performance may differ as well. I tried several tube brands and the JJ ECC83S and ECC803S seemed to be the best commonly available tube in this application. Do not use the new production tubes that start with an "M" and are claimed to be quiet as the S/N with them is about 3 dB worse than with the JJ brand tubes. Perhaps they work in other applications, but not here.
UPDATE - March 2016
As projects age I find that updates may be needed and in some cases can make a significant difference in how the project functions. This update covers one error fix, describes a possible performance change and gives details on an often needed improvement that can lower the noise floor of the phono preamplifier by a few dB. Please read the updates below in conjunction with the updated March 2016 schematic shown in Figure 5 (Click schematic images to enlarge).
The error fix is in the power supply. The resistors that "program" the LR8 regulator in the earlier schematics are not correct. Odd as it may be they actually work in one of my prototypes. They should however be 1k and 180k. This will provide 210-215 VDC to the preamp.
Figure 5: Updated Groovewatt Tube RIAA Phono Preamp Schematic (March 2016)
The possible performance change involves the coupling capacitor following the first stage. It is shown as a 0.22 uF. This value will provide a -3 dB point in the 40 Hz range. This can be beneficial for several reasons. First, it will eliminate problems with tonearm / cartridge resonances. Second it can reduce rumble that some records and turntables exhibit. Third, even though the preamp is very quiet, the noise it does have is higher in the 20 Hz and lower range so the 40 Hz slope makes it a bit quieter. Finally it can ease the load on some power amplifiers and speaker systems that have difficulty in reproducing very low bass. There is hardly any musical content in the sub 40 Hz range (yes I know there are some exceptions). I find that comparing two preamps using the differing values that it is very seldom audible. My overall system has an F3 of 18 Hz from end to end so it would be pretty obvious if there was a noteworthy deficiency. Use which ever value suits you.
The potential improvement is the addition of an electrolytic capacitor between the heater circuit and the signal ground. I have mentioned this before in the forum. When it works it can reduce the background noise level about 2 db. Since I hate noise this is a good thing. The problem is that it seems build specific. Sometimes it works and sometimes not. Worse sometimes it works better on the more positive side of the heater circuit and sometimes on the less positive. I can find no consistent pattern. What I do is use the small terminal pins (like in a PC) and put a jumper on the way it works best. If neither works then I leave the jumper off.
About the Author
Bruce Heran is the VP for design and support at Oddwatt Audio. He has been involved in the field of electronics in one form or another for nearly 50 years. Bruce is a strong supporter of DIY audio projects community and generally specializes in vacuum tube designs for Hi-Fi use. One of his main goals is to educate and encourage new comers in the building of affordable high performance audio equipment.
