Hoffman Amplifiers Tube Amplifier Forum
Amp Stuff => Tube Amp Building - Tweaks - Repairs => Topic started by: TubeMageinTraining on May 06, 2021, 05:21:06 pm
-
I just finished building a nice little 2 watt amp from a 1952 Travler 5170 clock radio. After using my magic shoe horn to get an isolation transformer and properly impedanced output transformer for an 8 ohm external speaker in there, it was all downhill. Well sort of. There were a lot of firsts...for me, but it went excellent & I learned a lot. I used the radio's 12BA6 IF stage pentode for one input and its 12AV6 detector/AF pre-amp triode for a second input. Fantastic amp for clean work. If anyone is interested, I can post sound samples. Triode is best for clean tele to rockabilly crunch and the pentode was just heaven for acoustic guitar. I used the pentode's data sheet and graphs to calculate design values and came up with what appeared to be a sound design. Then, I began researching circuits similar to what I had come up with to cross check my work with the experts of old. What I found as the closest examples with available YouTube sound demos were a Vox Student V5 and a Kay 703c. Both in the general "Widow Maker" series filament class of amps just like the AA5 radio I was converting. Looking at the component values of those amps and trying to apply the data sheets to validate those designs, I discovered the bias currents were anemic and the load lines were way in the basement off the charts. My design values were vastly different. I confess due to time constraints to deliver the amp, I caved and applied the Vox circuit values as the 12AU6 characteristics are very similar to my tube because the demos I heard were just what I was shooting for. I'm very happy with the results, but I'm still in the dark regarding designing for great pre-amp pentode tones, clean primarily. I am planning a "tube warmer" mic preamp project for studio work. The plan is one input into 3 tubes - 3 signal paths 1) a pentode likely a 6SJ7, 2) a dual triode 12AY7; 1 cathode biased triode and 1 AC cathode follower and 3) a DC coupled cathode follower each path selected for different non OD harmonic character to be mixed to a single output for a variety of harmonic colors. One of the holy grails of vintage tube preamps is an RCA OP-6. There are those pentode stages with crazy off the charts design values again. It looks like they are running the tubes way in the curve near cutoff. I need to get to the bottom of this. Can anyone help?
-
So, it occurs to me last night that with the load lines off the published chart, something needs to be done with the chart. Because I built the Vox Student V5 pentode preamp into my design, I had the quiescent voltages recorded. I'm sure I would have known this if I had read a book but since my G2 voltage measured 8V and the chart had curves recorded with the G2 at 100 volts, I just multiplied the G1 curve values by my G2 voltage ratio of .08 and did the same with the Anode current scale. With this new scaling, I was able to apply a load line and bias that jives with the design values. I've attached an image of my result. Is this how this normally done? If so, how far can one extrapolate like this and get reasonable results? One other thing... With a grid leak bias design like this, I presume the bias slides down the load line and biases cooler as the input cap charges, yes? If so how far and how much should one compensate to keep operation in the targeted region?
-
"Child's Law"
You ideally use the 3/2 power of voltage.
But you have a working chassis. Just dink with it until it works well.
-
So, it occurs to me last night that with the load lines off the published chart, something needs to be done with the chart. Because I built the Vox Student V5 pentode preamp into my design, I had the quiescent voltages recorded. I'm sure I would have known this if I had read a book but since my G2 voltage measured 8V and the chart had curves recorded with the G2 at 100 volts, I just multiplied the G1 curve values by my G2 voltage ratio of .08 and did the same with the Anode current scale. With this new scaling, I was able to apply a load line and bias that jives with the design values. I've attached an image of my result. Is this how this normally done? If so, how far can one extrapolate like this and get reasonable results? One other thing... With a grid leak bias design like this, I presume the bias slides down the load line and biases cooler as the input cap charges, yes? If so how far and how much should one compensate to keep operation in the targeted region?
So was your HT supply for the pentode only 120V?
-
Yes. It was only 120 as I was converting an old All American 5 radio with series heaters "widowmaker type circuit. I get a message from my brother, " this is a tube alarm clock radio inside dimensions 10.75x4.75x4.75. can you make it into an amplifier? Call me." The thing is the size of a box of tissue and includes a clock movement. See photos. I just put in a Triad 115:115 isolation transformer and went with it. The attachments tell the story best. Very basic schematic with a few things I wanted to try. Original radio OT transformer primary was re-used as a choke and I used the variable tuning cap as a tone control.
-
Nice job!