Cranberry TwoC - Cranberry All Tube Stereo Amplifier
Bruce Heran  
|
Cranberry TwoC - Cranberry All Tube Stereo Amplifier
The Cranberry is a stereo integrated all tube amplifier. I picked the name because of the paint color. It didn't come out exactly as I intended but looks fine anyhow. It is a heavy amplifier at nearly 18 kg and thus the need for handles on it. I suggest you use a steel chassis as aluminum will tend to bend under the stress. The amplifier uses proven power output stages like all Oddblock amplifiers and has an extra gain stage at the input to allow for use without a line stage preamplifier. Added features are the use of a balance control and input selection. Input sensitivity is approximately 0.1 Volts for full power output. The output power of the amplifier is determined by the choice of power tubes. Any pentode or beam power tube with voltage ratings of more than 500 Volts can be used. This includes ones like 6L6GC, EL34, KT77, KT88, and even KT120. All similar families of tubes can be used as well. The key parameters other than voltage rating are dissipation and current capability. I list some suggested tube choices and values at the end. The most important rating is likely the dissipation. To ensure good tube life I recommend you do not exceed 80% of the tube's rating. As in previous designs, the amplifier uses a relatively high output load impedance. This is a trade-off for better sound in lieu of maximum power output. It is also a class "A" design and as such runs the output tubes at maximum dissipation at zero signal. They will run hot.
Photograph 1: Cranberry TwoC - Cranberry All Tube Stereo Amplifier
Cranberry All Tube Stereo Amplifier Details
I will not discuss the power output stage operation much as it is covered extensively in my other posted OddWatt projects (see OddBlocks - Class-A Push-Pull KT88 Tube Amplifiers and also the KT77 Push-Pull Class-A Ultra-Linear Tube Amp). The amplifier circuit is a self-inverting push-pull (SIPP) output stage with a constant current source in the cathodes. No bias adjustments are needed and tube balance is controlled by shifting the individual cathode voltages slightly. This effectively changes the individual grid to cathode voltages as the grids are referenced to the signal ground. It looks odd (hence the name Oddblock), but works really well.
Figure 1: Cranberry All Tube Stereo Amplifier Schematic
Click Schematic to Enlarge.
The input stage is a SRPP design with modest gain of about 7. For folks that need more gain substituting a 6/12SL7 for the 6/12SN7 with corresponding circuit changes will provide a gain of about 35. It will increase the noise floor by a similar percentage though. Using the 6/12SN7 provides enough gain so that sources like a MP3 player or tablet will be sufficient for full output. BTW the reason for the dual tube types is that I use 12 VDC heater circuits as it is easier to obtain than 6 VDC at twice the current. Either will work fine though. An additional benefit is that there are some really nice NOS 12SL7s and 12SN7s that are far less costly than the 6 Volt versions. The amplifier uses capacitive multiplier circuitry in place of regulation (like earlier Oddwatts) for the driver and preamp stage high voltages. The thing to be aware of is the maximum voltage applied to the tube anodes. It really must be below about 325. Any more than that will eventually cause a heater to cathode failure of the driver or preamp tubes. If your power transformer delivers voltages different from the ones in the schematic then you will need to adjust the values of the resistors in the high voltage string. If your voltages are lower that is fine down to about 250 VDC in place of the indicated 315. It will have no effect on the sound or output power. Voltages below that may reduce the drive to the power tubes slightly and cause a slight loss of output power. In my personal build I used a thermal delay relay the Amperex 6NO60 to provide a B+ delay on start up. It triggers a relay to link the rectifiers to the filters. At start up there is no high voltage on the filters. Since the delay tubes are getting scarce and costly I have provided a delay circuit based on the LM555 IC timer that can be used in place of the delay tube. Alternatively you can just use a switch. I don't like that as it is too easy to forget and have the full high voltage dumped on the filters (and the SRPP stages) before any tube conducts. That voltage will be well over the ratings of several components and will probably cause circuit damage.
Figure 2: Alternate Delay Circuit with LM555 - Schematic
The output of the preamplifier stage is fed to a balance control and volume control. This is a compromise of sorts. Putting them there isolates them from the signal source and prevents loading. It also sort of enhances the end output signal to noise level. This is because any noise generated by the input stage is reduced by the setting on the volume control. The tradeoff is that it is possible to overload the input stage with very large signals. If you look at the schematic I suggest a number of different input series resistors. These resistors act as a signal divider with the grid resistor of the input stage and will ensure that the stage is not overdriven. Since the stage can swing over 50 Volts of output signal it may be a moot point, but I like to keep output signals from such stages below about 15-20 Volts to ensure maximum linearity. The input stage has no coupling capacitor and in my experience none is usually needed as virtually all signal sources do not have significant levels of DC voltage offset. If you encounter a source that differs from this I would add a low voltage poly capacitor of about 1 uf between the input resistors and tube grid.
Figure 3: Cranberry All Tube Stereo Amp Power Supply Schematic
Click Schematic to Enlarge.
Warning: This project uses potentially lethal voltages and should not be undertaken by anyone who is not familiar with working with such voltages or may not be comfortable with projects that entail such voltages.
Building the Cranberry All Tube Stereo Amplifier
The build is relatively uncomplicated. I like to spread out components when possible. The lead length and spacing is not a big issue in circuity like this that has only modest gain. It becomes critical in high gain circuits like phonograph preamps though. I recommend using shielded cable when a signal source has to travel more than about 3 inches such as between the input jacks and selector switch. Be sure to ground only one end unless it is actually used as a return conductor to avoid ground loops. Allow for adequate ventilation for the LM317s, the heater circuit rectifiers, power resistors in the high voltage section and the IRF820s in the capacitor multipliers. I like to put the LM317s on heatsinks that stick up through the chassis. They will need to dissipate between 5 and 10 Watts depending on your choice of tubes and will quickly cook things inside of the chassis if they are inside it. I tend to build things in a semi modular manner with a number of the circuit sections on small PCBs. How you actually wire or layout your build is pretty much of no consequence. I highly recommend that you make it easy to troubleshoot as stuff happens. I also recommend that you use a chassis at least 50% bigger than you think you need. Cramming stuff in a too small chassis IMO always causes problems. For wiring I like to use 20 and 22 gauge solid wire. You can bend it and it stays bent looking nice. Stranded wire IMO sometimes is not as good to strip and solder. It is fairly easy to cut off some stands reducing the effective size and may not always have good solder flow internally. I recommend using high heat (775F) for soldering as it will do the job well in a short time. I find that using lower heat requires longer time and can damage some sensitive components.
Photograph 2: Underside - Cranberry All Tube Stereo Amplifier
Photograph 3: Power Supply - Cranberry All Tube Stereo Amplifier
For additional valve amplifier design and construction tips, see my blog entry about design and construction tips and suggestions for vacuum tube amplifiers. Also, I have also posted some suggestions for a tube amplifier wiring color code.
Photograph 4: Capacitor Multiplier
Cranberry Tube Amplifier Operation and Final Checks
Initial startup and testing is fairly simple. One common mistake that DIYers make fairly often is to test the voltages on the power supply without having tubes in the circuit. The values measured this way will be quite a bit higher than the ones indicated in the schematic. They are only valid when there are tubes in place. The high voltage values measured without tubes can easily exceed the voltage rating of the filter capacitors and may cause rapid failure. With the tubes in place the values can vary a bit. It depends on your actual AC mains voltage, the power transformer you used and the choice of output tubes. Less powerful tubes will pull down the high voltage less than the largest ones and will thus cause an increase in the actual high voltage. The typical range with the transformers listed is between 425 for the largest tubes to 475 for the less powerful ones. As mentioned earlier check the voltages on the drivers and preamp tubes to be sure they are below 325 and adjust the series filter resistors to obtain the correct values. A word for new DIYers in the tube field. They are not dainty devices and while should never run them above their specifications, they will for a while tolerate voltages and currents well over their ratings. So if your initial startup results in excess voltages or current through the tubes when measured at the cathode test points do not be alarmed. Just shut it down and hunt for the cause. A minute or two will not normally cause the tubes problems.
The startup procedure is simple and if you have two inexpensive meters super easy. I have a batch of "give away" ones from Harbor Freight. They only retail for about $7US and with a coupon often are free. They are not the best ever made, but work fine for initial checking of the tube cathodes. The reason for two meters is that there are two channels and you really want to know that both are functioning. I like to place a load on the amplifier output typically a modes speaker to hear if anything occurs on power up. The amplifiers are sufficiently stable that it is optional, but it is a good verification that the amp is not producing oscillations or loud hum. Set the meter(s) for 0-1 VDC. Hook one up to one set of test points for each pair of tubes. What you want to see is something in the 50 to 150 millivolt range. Zero or near zero means that particular tube is not conducting for some reason. Immediately shut off the power. A common reason for this is a wiring error. The most serious is that the grid circuit on that tube is not grounded correctly and the cathode voltage went way high and fried the LM317. Other problems are connections to the high voltage power source that are not powering the tube. Since there already is an error this is one time that you will need to measure the high voltage. Do so at the anode of the non-conducting tube. If it is high then the problem is either a failed LM317, wiring error in the cathode circuit or very remotely a failed tube. If the value you measured is over about 200mv then that tube is hogging all the power. Look for errors in the grid circuits. If the voltage across the test points is way high like 300-400mv then the LM317 has failed shorted through or else you have somehow bypassed it in the circuit. If the value is about what you expected based on the tube type used (mostly in the 60mv range for low power tubes and 90mv for the bigger ones) then you can assume that the particular pair in functioning OK. If all 4 tubes are in the right range then quickly verify that the voltages on the driver and preamp tube upper anodes are between 250 and 325 Volts. If not look for the source of the difference. Now go back to the power tubes and attempt to balance the readings on the test points using the control. The values for each pair should be similar but allowing for parts tolerances unlikely to be exactly the same. Allow for some time to fully warm up and check them again. Repeat this after a few hours of use and at about one month. After that they ought to stay sufficiently in range for about 6 months. A variation of about 5-10% is fine over time and does not impact the sound or performance in any significant way. Much more than that indicated that tubes in that pair are not well enough matched to be used. Actually matched pairs are not required in this amplifier, but tubes should be from the same batch. Mis-matched tubes will not balance correctly and likely the one passing the most current will fail. If the voltages and currents are in the proper range then it is time to hook up speakers (if you didn't do so earlier) and feed in a sound source. The output should be clear and undistorted. If not there is some type of error. If it is only both channels it is likely related to a power supply problem or grounding issue. If only one channel you can check voltages and resistances (power off and amp idle for several minutes) and compare one side to the other looking for differences. If the overall sound is thin and seems to not have enough bass then it is likely the phase of the output transformer(s) is backwards. They are used to invert the signal phase to make the amp non-inverting and have the really minor amount of negative feedback have the correct phase. If this is the case you have a few options. Pick the one that is easiest. Swap the connections on the grids to the power tubes. That would effectively swap the active tube with the one that is slaved to it. Alternatively you can swap the wires from each anode and screen from one tube to the other. The third option is you can swap the ground and hot output leads on the output transformers. This only works if you use only one impedance. At this point the amp should be ready for use. A final word on this is that the common thought that you should re-tube amps regularly does not apply to these amps. I am still using the original KT88s from my first big amp. It has now been nearly 13 years and many thousands of hours. They sound and test fine.
Suggested output stage tubes.
High power (92 mA per with output power in the 28-33 Watt range)
- KT88
- KT90
- Any power pentode that can handle 500 VDC and dissipate over 40 Watts
Lower power (62 mA per tube with output power in the 15-18 Watt range)
- EL34
- KT77
- 6L6GC (new production tubes that can handle the dissipation)
- Any power pentode that can handle 500 VDC and dissipate over 30-35 Watts
Other Tubes
KT120 running as KT88 at 92 mA and 28-33 Watts output or my favorite as KT77 at 62 mA and 19-20 Watts output. The reason this setting is my favorite as to me they sound the best of all combinations.
Use and Listening
Be sure to place the amplifier where is can breathe. It will get really hot. Class A amplifiers are like that. I recommend speaker with relatively high sensitivities. For the lower rated tubes something in the 90-92 dB/Watt is typical. For the KT88 and similar versions the sensitivity can be lower. It all depends on how loud you want the sound to be and how large the listening room is.
About the Author
Bruce Heran is the VP for Design and Support of Oddwatt Audio. He has been involved in the field of electronics in one form or another for nearly 50 years specializing in Hi-Fi vacuum tube designs. One of his main goals is to educate and encourage new comers in the building of affordable high performance audio equipment.