Audio Entity

Audio Entity

Audio Entity Blog

high voltage hurts

This is where information about project progress plus vacuum tube amplification related circuits and other tidbits of interest will be posted. The Blog is not intended to be a forum, no method of leaving comments is provided. The Author or Site Administrator or Audio Entity are not responsible for any harm or damage, direct or indirect, from use of information contained in this blog or on Vacuum tube circuits use voltage and currents that can be lethal. As with all electronic work steps should be taken to work carefully at all times.

Blog created 06/09/2020

Update 08/01/2020

Normally I do not use turret pin strips. They tend to make wiring more spread out and more prone to picking up hum or other noise. In this case, I thought they would be a good choice for the base chassis since there are no audio circuits. The top panel uses a combination of terminal strips and turret tie points.

What I am finding out is the turret pins do not take solder very well. The pins are silver in color, I'm not sure if they are tinned or just plated. Switching to a hotter soldering tip helps and once the solder grabs it flows nicely. The problem seems to be the initial contact between the solder and turret pin. I imagine these turret strips are RoHS compliant meaning they probably don't have any decent tinning on them.

Blog created 08/01/2020

Update 07/28/2020

6SN7 Chassis Base

After several days of careful component placement, drilling and a couple corrections the chassis base is ready to wire. Wiring will take several days of determining wire lengths for the harness between the chassis base and top panel plus testing.

Except for leads coming out of transformers or other components the chassis base (and top panel) will be wired using PTFE (Teflon) insulated stranded wire.

6SN7 Chassis Base

In order to solder tie point strips some components will need to be removed during soldering to help prevent melting something with the hot soldering iron barrel.

Blog created 07/28/2020

Update 07/21/2020


There was a delay getting the top panel made, it arrived yesterday. The panel will mount on top of the chassis base and act as the amplifier chassis top.

Layout of the panel will include a blue EDOR power transformer in the center of the panel. There is a hole to the left and right of the power transformer for posts to help support weight of the transformer. The two sets of three holes in a triangle pattern are the inputs and outputs. Holes right below channel 1 and channel 2 are for RCA input jacks. The two holes between the input jacks and power transformer are for speaker connection posts.

There are gain trim controls on either side of the volume control. Besides balancing the volume between channels the trim controls allow adjusting gain of the amplifier so the volume control works in a comfortable range. The two big holes are for meters to monitor cathode current of the output tubes, one meter for each channel. Power on indicator is to the left of the power switch. A bias failure indicator lights up should either channel bias supply fail.

Next step is to finish drilling holes in the chassis base. The top panel will be used to mark off, drill and tap holes for mounting the top panel to the chassis base.

NOTE: Handles did not come with the panel, they were installed for the picture.

Blog created 07/21/2020

Update 06/26

Started drilling holes for components. The base is shown with output transformers, high voltage rectifier board and bias supplies temporarily mounted. What you are looking at is a painted Hammond chassis turned upside down and will be the amplifier base. A 1/8" thick engraved panel is being machined to fit on top of the chassis. The panel will have the tubes, power transformer, input and output connectors, controls and cathode current meters. The panel has not yet arrived from the vendor. Base fabrication has to wait for the panel to arrive.

Blog created 06/26/2020

Update 06/24

Mounted components on the high voltage printed circuit board and tested. The PCB has rectifiers rated at 1500 volts at 1 amp, 6 - 33uF 400 volt filter caps connected in series/parallel and two 100K 3 watt resistors to balance voltage across the parallel capacitors. The result is a capacitor rated about 50uF at 800 volts. The 100K resistors also serve as a 200K safety bleeder resistor.

The high voltage board was created to replace axial 600V capacitors that have become expensive and hard to find. As a convenience rectifier diodes are also on the board.

Some folks prefer a vacuum tube rectifier over solid state. A tube rectifier holds off high voltage until the rectifier cathode warms up. This does prevent higher than normal B+ voltage before the amplifier tubes warm up and start drawing current. The downside of using a tube rectifier is the high internal resistance of the tube limiting current. This causes the B+ voltage to fluctuate quite a bit depending on power output current demands at different volume levels. The voltage drop across solid state rectifiers is negligible providing a more solid B+ voltage. Since this design is using a bias system that mimics push-pull current demands a more solid B+ voltage is preferred. As far as the B+ being higher until amplifying tubes warm up, the tubes will not start amplifying until they warm up reducing the likelihood of audio plate voltage peaks exceeding safe operation.

Blog created 06/24/2020

Update 06/22

Just about all supplies needed to build a prototype are on hand. Waiting for a price to have an engraved panel made plus a few odds and ends to be delivered.

Have to wait until I have the engraved panel before I can lay out component mounting inside the chassis. The engraved panel serves as the top of the amplifier with tube sockets, power transformer, cathode current meters and controls. Everything else is inside the chassis base. Components and wiring on the top chassis panel have to clear components inside the chassis base. One cable harness connects the two together. The harness will be long enough to allow the top panel to be off the base chassis and still be able to power the amplifier up for testing.

Blog created 06/22/2020

Update 06/15

Delays to relocate look indefinite for now. Unpacked a few boxes to start work on the responsive bias project.

At this point experimenting with component layout inside the chassis. The blue power transformer mounts on top of the chassis, under the power transformer inside the chassis is where the filter choke goes.

It's a dual channel stereo Class A amplifier using triode output tubes. The output tubes will be four 6SN7's connected in parallel. Power output should be about twelve watts per channel or thereabouts. Early experiments using two 6SN7's in parallel provided around 4 watts and produced a wonderful rich sound. The driver tube is also a 6SN7.

Bias control boards will mount to the left and right sides. Output transformers are mounted inside the chassis at the left and right backside. They are rated at 20 watts. The PC board next to the power transformer will contain the rectifier diodes and first filtering. The capacitors are small, but have a high ripple current rating and are long life. They are connected series/parallel for a capacitance and voltage rating of about 50uF and 800 volts. Voltage balancing resistors on the board also serve as safety bleeders

The only components or hardware visible on the top side of the chassis will be the power transformer, tubes and two handles for lifting the amplifier. This should make for a clean uncluttered appearance. The tubes will remain exposed without any protective covers. The 6SN7's do get hot, but not to the point of causing a second or third degree burn.

Blog created 06/15/2020

Responsive Bias Status

As of today June 12th covid-19 virus restrictions still apply in the Philippines. My wife and two young daughters are presently staying with her parents. We plan to live in Davao 140 miles away, but travel restrictions have prevented her from searching for a place to live. Project is on hold and everything associated with the project including test equipment and tools are packed waiting to be shipped. Problem is until she finds a place to live I do not have an address to ship to. Even when she is able to find a place shipping by sea freight takes about three months.

Blog created 06/12/2020

Parallel Class A Output

Two output tubes in push-pull equals two output tubes in parallel. To some extent this is true. Power output from two tubes in push-pull is the same as two tubes in parallel. There are some differences in performance.

Push-pull usually operates Class AB where each output tube cuts off for a portion of the signal waveform. This reduces current drawn by the tubes and less power is wasted as heat. Tubes connected in parallel operate Class A where both tubes operate totally linear. Class A draws more current, but not twice as much. Class A operation is biased in the middle of its operating curve. From the middle current swings up and down the operating curve as it follows the signal waveform.

Class A operates completely linear, there is no manipulating of the amplified signal. Class AB splits the signal to supply the positive portion of signal waveform to one output tube and the negative portion to the other output tube. This sort of manipulation generates undesirable artifacts. Class AB cancels the most desired second harmonic, but leaves the least desired third harmonic distortion content. Both halves of the split signal must be amplified exactly the same or else the regenerated signal in the output stage will have added distortion. Even if component values are a tight tolerance there will be some difference between signal halves.

Class A pentode parallel output
Class A pentode parallel output

The circuit shows two pentode output tubes connected in parallel operating Class A ultra-linear with fixed bias. Most of the components are typical. For added protection against parasitic oscillation R1 and R2 are in the plate circuit. R3, R4, R5 and R6 also serve to limit parasitic oscillation. The value of R7 is equal to the value of the output tube grid resistor value, 100K to 150K or as specified in tube datasheets. The cathode resistor R8 provides a small voltage drop at the cathode to provide a cathode current reference voltage at test point TP. The test point can be eliminated if R8 is replaced with a current meter to monitor cathode current.

R1 & R2 - 10 3W Metal Film
R3 & R4 - 150 3W Metal Film
R5 & R6 - 10K ½W Metal Film
R7 - (Read Text) ½W
R8 - 20 5W Wirewound
C1 - .47uF or larger

When operating tubes in parallel Class A make sure you do not exceed maximum plate dissipation. Plate voltage times plate current is power dissipated by the plate.

R1 and R2 was lowered to 10 ohms to account for lower load resistance. For load resistance of 5K or above use 100 ohms. Load resistance for pentodes is plate voltage divided by the plate current.

Blog created 06/11/2020

Load Effect on Supply Filtering

loading effect on power supplies

Multiplying the desired rectifier output DC voltage by .707 will give you the transformer secondary voltage under no load conditions. For example, 400VDC X .707 would require a 566VCT (283-0-283) secondary. Since capacitors charge to the peak value of the rectified voltage the no load DC voltage would be 283V X 1.414 = 400VDC.

loading effect on power supplies

Any load placed onto a DC supply is also a load placed on the supply capacitors discharging them between AC pulses. The result is a DC voltage less than the peak value. It is possible to compensate by using a power transformer rated higher voltage. You can approximate how much more voltage by factoring a 19% load loss. To calculate approximately how much more voltage you need, multiply .19 times the no load value. Add the result to the no load value. For example, 450VDC at 200mA (450 X .707 = 318) (318 X .19 = 60, 318 + 60 = 378). A close match would be a transformer rated 760VCT (380-0-380) @ 200mA.

Keep in mind that you should account for voltage drop across resistance in the filtering circuit such as a filter choke or resistor.

Blog created 06/10/2020, updated 07/18/2020

Transformerless Power Supply

power supply without power transformer

In 1955 at the age of 10 I started learning about electronics using vacuum tubes. At the time there was a popular power supply circuit that did not use a power transformer. It used a half wave rectifier powered directly from an AC outlet. This transformerless circuit was primarily used in table radios and low power audio amplifiers.

There were several tube types specifically designed to be used with this type of power supply. Filaments were connected in series and connected directly across the AC power. The total of all the rated filament voltages had to equal the AC power, in the US 110V to 120V. Filaments had to all have the same current rating of .15 amps. The same current rating would divide 120VAC between the various filaments with the proper voltage at each filament. The 35W4 rectifier in the drawing had a filament rated 35 volts at .15 amps. A 50C5 output tube had a 50 volt filament rated .15 amps and was available as a 35C5 with a 35 volt .15 amp filament. If the total of the filament voltage values was less than 110V to 120V, then a series resistance was added to the series filament circuit. A few current popular tubes are left over from the series filament era. This includes 12AX7, 12AT7 and 12AU7 tubes all having a 12 volt at .15 amp filament.

power supply without power transformer

Without a power transformer for isolation the chassis cannot be connected directly to ground. Instead the chassis was grounded through a high value resistance with a .05uF to .1uF capacitor across the resistance. The high resistance limited current to prevent a dangerous shock and the capacitor provided an audio ground path for input jacks.

There are obvious deficiencies with this type of power supply. AC outlets back then were not polarized, plugs could be inserted either way into an outlet. Depending which way a plug was in an outlet the ground circuit might actually be on the hot AC leg and would cause a hum in the audio. If one tube has a filament open up, then all the tubes lose filament power. Without a transformer to boost high voltage the most B+ voltage you could get was 150VDC to 160VDC depending on the B+ load. Despite its deficiencies the circuit worked fairly well for its intended purpose.

Blog created 06/10/2020

6SN7 High-Level Driver

6SN7 high-level driver stage

A 6SN7 makes an excellent driver stage to drive single ended output tubes. It is also capable of providing some fairly high grid drive voltage. In order to provide high-level output more B+ supply voltage is required. More B+ voltage gives the tube more voltage to swing and extends the linear portion of the operating curve. Using a B+ supply voltage of +450 to +500 volts will result in voltage across the 6SN7 to be 250V or less, well below the 450V plate voltage maximum.

6SN7 high-level driver stage

The circuit uses both sections of a 6SN7 as a single driver stage. Plate to plate negative feedback reduces distortion and extends frequency response. In this application negative feedback does not produce undesirable artifacts. This is because under normal operating conditions output of the driver should be well below clipping and not introduce distortion back to the input stage. Also, with feedback back to the previous stage plate instead of the cathode results in less tracking error in the feedback loop. This driver would work well to drive a KT88 single ended Class A output where a +500 volt plate supply is used.

B+ supply voltage = +500V
Plate current = 12mA (6mA per section)
Maximum output before clipping = 100VRMS
−1dB points 7HZ & 55KHZ
Voltage Gain 36

Output = 30VRMS (42 volts peak)
Distortion 20HZ = 0.12%
Distortion 100HZ = 0.10%
Distortion 1KHZ = 0.20%
Distortion 10KHZ = 0.32%
Distortion 20KHZ = 0.47%

Output = 65VRMS (92 volts peak)
Distortion 20HZ = 0.46%
Distortion 100HZ = 0.46%
Distortion 1KHZ = 0.52%
Distortion 10KHZ = 0.80%
Distortion 20KHZ = 1.40%

Measurements made on a HP331A analyzer

When considering harmonic distortion remember that harmonics of 10KHZ start at 20KHZ, about the upper limit for humans. Harmonics of 20KHZ start at 40KHZ.

Blog created 06/09/2020

Audio Entity Main