uTracer (Micro Curve Tracer)

General discussion area for tube amps.

Moderators: pompeiisneaks, Colossal

Post Reply
User avatar
martin manning
Posts: 13080
Joined: Sun Jul 06, 2008 12:43 am
Location: 39°06' N 84°30' W

Re: uTracer (Micro Curve Tracer)

Post by martin manning »

Subminiature tubes... how to test them with the uTracer? No problem! A small breadboard and some banana plugs with short lengths of wire worked nicely for this 6205 pentode, but not without a ferrite on the g1 lead to stop it from oscillating. An initial QuickTest (without the ferrite) produced odd results, with plate current much higher than expected. Tracing the anode curves revealed the problem. This tube has a separate pin for g3 (like EL34's) which is connected to k with a jumper.
6205-uTracer-test.png
Shown here are the curves as traced without the ferrite. Previously I've seen curves that appear crumpled, but in this case Ia is seen to take off above -2.5V Vg1 at some Va. Unfortunately, the data sheet spec point (Va = 100, Vg2 = 100, Vg1 = -1.5V) is in the unstable region. I wonder if this is part of the reason vintage testers use low voltage?
6205_Trace_Osc.PNG
Here's a nice clean trace, run with the ferrite added to the g1 lead:
6205_Trace_Stable.PNG
And here is the QT result, showing that the tube is close to spec in Ia, Ig2, and gm. Ra is quite low, but that parameter is both difficult to measure and not important if you have the complete curve.
6205_QT_Stable1.PNG
Here I've taken the opportunity to illustrate a uTracer quirk, which is that the plate and screen voltages will sag a few volts below the set point during the measurement pulse. Fortunately, the actual voltage that appears during the test pulse is captured on the QuickTest form, so the set point can be adjusted upward to hit the desired target point. In this case, I had to set 103.4V. to get 99.8V, targeting 100V Vg2. This is a great benefit of the QuickTest function, since the characteristics at an exact point can be measured and compared against data sheet values. In the plot below the full trace is run with nominal voltage (100V) set for Vg2. Then, a QuickTest was run to find the adjusted voltages. Following that, I clicked back to the main form and ticked the "Keep Plot" box before clicking in the plot area to superimpose the QuickTest data using the adjusted voltages on the nominal Vg2 curves. I did this twice, at the QT spec point and at the same Va with Vg1 set to 0, where the Vg2 adjustment required was approximately the same (within a couple of tenths of a volt). Finally, I increased the Vg2 set point on the main form and reran the Vg1 = -1.5 and Vg1 = 0 curves (shown in black), which have now shifted up to coincide with the QT data. Exactly how much the screen voltage sags depends upon how much current is drawn, so in point of detail the required offset is not a constant. Here that makes little difference. In all cases, Va is plotted at the actual test voltage, so no adjustment is needed there. This little dance is required if you want to plot accurate pentode curves.
6205_Trace_QT_Vs_Corr_Vg2.bmp
You do not have the required permissions to view the files attached to this post.
Last edited by martin manning on Wed Nov 08, 2017 11:49 am, edited 5 times in total.
User avatar
M Fowler
Posts: 14017
Joined: Wed Jan 16, 2008 2:19 am
Location: Walcott ND

Re: uTracer (Micro Curve Tracer)

Post by M Fowler »

Your quite the mad scientist Martin, love your U-tracer, wish I was smart enough to build one let alone know how to use the darn thing.
Bravo :lol:
Last edited by M Fowler on Sun Nov 05, 2017 2:51 pm, edited 1 time in total.
User avatar
martin manning
Posts: 13080
Joined: Sun Jul 06, 2008 12:43 am
Location: 39°06' N 84°30' W

Re: uTracer (Micro Curve Tracer)

Post by martin manning »

I have no doubt that you could build and use one, Mark. Several issues have been worked out over time as you can see in this thread, and if you use a Windows machine loading and running the software is easy. IMO the only remaining problem with the basic device is the on-board heater supply. That can be managed, but it's just easier to use something else.
User avatar
martin manning
Posts: 13080
Joined: Sun Jul 06, 2008 12:43 am
Location: 39°06' N 84°30' W

Re: uTracer (Micro Curve Tracer)

Post by martin manning »

Here are two more measurements of the 6205 illustrating the effect of failing to compensate for voltage sag during testing. First, with the tube connected as a pentode, the anode and screen voltages were adjusted so that the test voltages were at the target values of 100V Va and 100V Vg2 in QuickTest. Those measurement results (Ia, Ig2, and Gm) were then entered as the nominal values in the QT form (establishing “100%” values for this tube at the target voltages). Then with the input voltages reset to the nominal values (100V), the test was run again. Those results are shown below, where it can be seen that the anode and screen currents have dropped 9% and 8%, respectively, and Gm is 4% lower. Ra and mu values from the first test were also entered, but those values will not be accurate for a pentode. The second QT form shows the result using the same tube and process, but this time with the tube connected as a triode. Here the anode current has dropped 12%, Gm is 4% lower, Ra is 5% higher, and mu is showing no change. Interestingly the screen voltage has sagged 3.6V during the test even though the screen supply is not connected. In both triode and pentode cases, these effects will show up in the traced curves, most noticeably in the plotted Ia and Ig2 levels as shown for this tube in the post from 4 November, above. Of course the need for calibration will apply to any tube tester, and in this case the plate current error can be more than twice the plate or screen voltage error, and the Gm error roughly equal to the voltage error, on a percentage basis.
You do not have the required permissions to view the files attached to this post.
Theodorus
Posts: 1
Joined: Wed Jan 01, 2014 12:08 am

Re: uTracer (Micro Curve Tracer)

Post by Theodorus »

Hi, I just finished my Utracer and am still learning how to work with it , used the idea for the panel from here (thanks for this) and made my own box .
The info in this forum has been very helpful , the unit is working fine so far i'm very pleased with it :)
At the moment i'm trying to make spice models with the date from the Utracer witch is a just lot of fun ..
Image
Image

Greetings Theo
User avatar
martin manning
Posts: 13080
Joined: Sun Jul 06, 2008 12:43 am
Location: 39°06' N 84°30' W

Re: uTracer (Micro Curve Tracer)

Post by martin manning »

Looks great! I've been very happy with that patch cord arrangement. A new version of the the GUI has just been released which accommodates the low voltage hardware developed last year, as well as a new "low grid bias" mode that uses some additional circuitry to create a -10 to +5V grid bias range (see uTracer site, lab notebook pages). This will solve the problem of setting accurate Vg1=0 using the current hardware as well as providing some positive Vg1 capability. I have designed a layout for that on a small daughter board, but Ronald has something further cooking so I'll wait a bit before I construct it. I'll post it here along with some test results when it's done.
User avatar
M Fowler
Posts: 14017
Joined: Wed Jan 16, 2008 2:19 am
Location: Walcott ND

Re: uTracer (Micro Curve Tracer)

Post by M Fowler »

Good looking u-tracer enclosure.
User avatar
sergio
Posts: 400
Joined: Mon Nov 26, 2007 9:36 pm
Location: italy
Contact:

Re: uTracer (Micro Curve Tracer)

Post by sergio »

Hi all, now I'm work with internal heater supply , but if I use the external supply , I have to insert some value in the Vh field and press "HEATER" before starting to measure? or the Vh field is irrilevant?

Thanks
S.
www.sae-audio.it
Handmade tube amplifiers inspired to Dumble sound!
User avatar
martin manning
Posts: 13080
Joined: Sun Jul 06, 2008 12:43 am
Location: 39°06' N 84°30' W

Re: uTracer (Micro Curve Tracer)

Post by martin manning »

First, you always have to press "Heater On" to start a measurement. If you use an external heater supply that is connected directly to the tube (my preference), then the voltage setting on the main form is irrelevant, and you do not check the external heater supply box. Note that the external supply must be floating if you want to trace tubes with directly heated cathodes, and see the above diagrams on directly heated rectifier tracing for details on the voltage offset that will result.

There is another option where you can connect an external heater supply through the uTracer PWM circuit, which is described in the uTracer user manual. This technique works for indirectly heated cathodes only, but it does allow automated heater voltage sweeps. You have to tick the external heater supply box on the main form to use this method.
User avatar
sergio
Posts: 400
Joined: Mon Nov 26, 2007 9:36 pm
Location: italy
Contact:

Re: uTracer (Micro Curve Tracer)

Post by sergio »

thanks Martin!
S.
www.sae-audio.it
Handmade tube amplifiers inspired to Dumble sound!
User avatar
martin manning
Posts: 13080
Joined: Sun Jul 06, 2008 12:43 am
Location: 39°06' N 84°30' W

Distortion Analysis

Post by martin manning »

Wow, 18 months since I've updated this thread. My uTracer has been down for some while due to a computer upgrade, which required a new memory partition (MacBook running Bootcamp), a new copy of Windows 10, and a new FTDI USB-TTL cable (the old one was obsoleted by the new Windows 10). It's all up and working again, and the latest GUI release has some options for additional hardware mods as well as some bug fixes. I haven't implemented the low/positive grid bias circuit yet, but I do have parts and a plan. Cramming that into my enclosure looks feasible, so that may yet happen.

Some discussion started in the Dumble forum around the choice of plate load resistance and the distortion produced. Dumble amps are seen to have a range of plate loads on the preamp stages, ranging from Fender-like 100k with 1k5 cathode resistor on all to several combinations of higher values up to a maximum of 220k with 3k3 cathode resistor. Merlin Blencowe gives a simple explanation of the effect of plate load on 2nd order harmonic distortion (pg. 22, here: http://www.valvewizard.co.uk/Common_Gain_Stage.pdf), but admits his analysis does not account for grid current. He shows 0.6% less HD2 for 220k vs. 100k Ra, assuming a constant Vgk bias point, and (evidently) that the signal level is a bit less than what would bring Vgk to zero on the positive swing. SPICE simulations show that the distortion profile is sensitive to both the bias point and the signal level.

The uTracer GUI can do distortion analysis, so let's see what that can add. The technique used is to spline-fit a set of plate curves and then analyze by FFT along a specified load line. Here I've traced a 12AX7, and used the distortion analysis to examine 100k and 220k load lines with a the same bias voltage (Vgk -1.55V, typical of 100k/1k5), and also with the Dumble high-plate 220k/3k3 combination, which shifts the bias point a bit cooler ( to -1.75V). In all three cases the signal level is set just below that which would take Vgk to zero, and also avoids driving the tube into cut-off.

The results at constant Vgk agree at least in direction with Blencowe, with HD2 being less for 220k by 1.1%. If the cooler bias point is used, the direction is reversed, with the 220k HD2 being higher by 0.3%. In HD2 dB down from the fundamental, the two are approximately the same, which agrees with the SPICE simulation using the cooler bias point, but the higher-order harmonics don't come up like they did in SPICE, leaving the THD about the same for the two values of Ra. At least it can be said that the uTracer analysis agrees that THD is sensitive to the bias point. Note that this analysis is along the DC load line, where the SPICE simulations have the coupling cap and following stage load on the anode. Tha makes the 220k look like 190k, and the 100k look like 91k.
You do not have the required permissions to view the files attached to this post.
Last edited by martin manning on Mon Sep 02, 2019 9:17 pm, edited 2 times in total.
User avatar
martin manning
Posts: 13080
Joined: Sun Jul 06, 2008 12:43 am
Location: 39°06' N 84°30' W

Distortion Analysis

Post by martin manning »

Example of 12AX7 curves. The load line is in green, with the portion used for the distortion analysis in black
You do not have the required permissions to view the files attached to this post.
samrock55
Posts: 64
Joined: Tue Sep 11, 2018 12:44 am
Location: State College PA

Re: uTracer (Micro Curve Tracer)

Post by samrock55 »

Heck just give me my Sencore Mighty 7 or an old Hickok anything beyond that put the tube in the Amp and listen.
User avatar
martin manning
Posts: 13080
Joined: Sun Jul 06, 2008 12:43 am
Location: 39°06' N 84°30' W

Re: uTracer (Micro Curve Tracer)

Post by martin manning »

samrock55 wrote: Tue Sep 10, 2019 5:49 am Heck just give me my Sencore Mighty 7 or an old Hickok anything beyond that put the tube in the Amp and listen.
Back on page two:
martin manning wrote: Mon Aug 19, 2013 1:19 am IMO vintage testers are like vintage cars: ok for a hobby and sunday driving, but limited in capability, not good for frequent use, and collectors are driving prices up.
User avatar
M Fowler
Posts: 14017
Joined: Wed Jan 16, 2008 2:19 am
Location: Walcott ND

Re: uTracer (Micro Curve Tracer)

Post by M Fowler »

Martin and others thanks for all the information you guys are providing. I bought a kit last year and waiting for cold winter and quiet times to get started on mine.

The only parts I need are those ferrite beads where did you get yours? I guess I need to know what size wire I will be using first.

Mark
Post Reply