k6xx at juno.com
k6xx at juno.com
Thu Jul 2 08:15:16 PDT 2009
Couple of comments.
Most importantly, the "wattmeter" in the AL-1200/1500/82 amps is NOT accurate. It is merely a voltmeter. SWR of any type (resistive or reactive) destroys any semblance of precision. I think it also is rather frequency sensitive, but won't make that claim.
You need an external wattmeter to have some assurance that your "1200W" is actual. On mine, I've seen 50% variation between the internal voltmeter and a Bird, differing from band to band as the antenna impedance changes. Even with a decent dummy load there is significant divergence. When I first got the AL-1200, I was 'astounded' to measure 82% plate efficiency on 20m! When it exceeded 100% at some test frequency, I was even more thrilled... but then a bit skeptical!
Second, confuzers and other IC-based devices may add temp sensors basically for "free", since a semiconductor diode junction is a decent thermometer. Hollow state amps--or anything in a big box--would need to mount descrete sensors around the box, so it would not be as low cost.
Here, I have metal dial thermometers mounted in a hole drilled through a small wooden block that sits atop the amp. The attached sensing probe is in the exhaust stream. I've empirically determined that when the temperature reaches about mid-dial reading, the amp is working well. If it is above half scale, something is wrong and requires correction.
This low tech solution uses thermometers purchased from Harbor Freight on sale for about $2 each. I bought a handful and the correlation between them is less than ideal, but you may rotate the thermometer in its mounting hole--so it may be "calibrated", sort-of! No flashing digits or lights, but it seems to do the job
73 de Bob, K6XX
Date: Mon, 29 Jun 2009 12:57:41 -0700
From: Kevin Normoyle <knormoyle at surfnetusa.com>
Subject: [Amps] AL1200, RTTY contesting, temps, duty cycles (not 100%)
To: AMPS at contesting.com
Message-ID: <4A491CB5.3050301 at surfnetusa.com>
Content-Type: text/plain; charset=ISO-8859-1; format=flowed
Looking for feedback. Dumping info and my understanding.
I got some remote temp sensors hooked up to a serial port (QK145 kit)
and wanted to investigate temps on my AL1200 during RTTY contesting.
Improved my dummy load so it could handle 1200 watts at 50% duty cycle
with a stable temp at the dummy load. (measured exhaust air on the dummy
load to see this).
As I started looking at this, it made me think that RTTY duty cycles are
not characterized correctly, in describing thermal requirements for an
amp. Key-down forever may be interesting as a upper bound, ignoring
cost, but saying RTTY contesting is comparable to key-down operation
seems wrong. RTTY rag-chew may get close to that, but I don't need RTTY
A lot of posts here talk about RTTY being a 100% duty cycle. But
nowadays, RTTY contesting has short exchanges. My worst case is a long
CQ, which is about 7 seconds. In aggressive times, I only wait 6 seconds
between. So that's a 50% duty cycle, with 7 seconds ON time. While
running , my anecdotal evidence is that the ON/OFF time is less severe
than this repeating CQ. (maybe 3 macros in 45 secs average case? so no
worse than 7 secs on, 7 secs off. 80/hr run rate?)
My reading about plate dissipation on tubes, makes me think that ON time
is still short enough, (for this tube) to not be talking about a 100%
duty cycle. This seems to make sense, because I don't read much about
So then, obviously, the issues must all be around long term heat rise.
And hot spots in the box. So the question is: what is the stable max
temp inside the box at any point, and how long does it take to get there
given my worst case repeating CQ?
So I ran the CQ repeating, for 30 minutes into my load and measured
exhaust air temp. Temp probe was about 4" above the exhaust outlet.
While this might sound stressful, it's no worse than real contest operation.
I limited the power to 1200 watts, as indicated on the AL1200. Grid
current was 180mA.
The dummy load is 4 globar-style resistors, air cooled. SWR 1.1:1. 80M
band. Plate voltage is 3500v, 3200v transmitting. Amp runs off 240v.
245v tap, with stock blower hooked to highest tap.
My thinking is that stable exhaust air temp is only reached when all
components in the box have reached a stable temp. Yes there are unknown
higher peak and avg temps inside the box. The temp sensors are Dallas
DS18S20. Stable was exhaust air temp staying within 0.5F over a 5 minute
I'm assuming some reasonable thermal conductivity from anywhere in the
box, to the exhaust air (whether through metal or air). If this wasn't
true, then even non-rtty activity would eventually cook something in the
Ambient air temp was 78F. Most of the exhaust temp rise happened in 5
minutes. Stable not until 15 minutes, at 147F. I would note that it took
5 minutes of idle with fans on, to reach a stable (lower: 103F)
non-transmitting temp. This is a good reminder to leave amp on for 5
minutes after transmitting, to cool down.
I realize that it is probably worse for the higher bands?
A couple thoughts/questions:
1) Ambient temperature control seems like the most direct way to control
amp temps (this amp). Also: having some airflow over the large black box
would seem to help box dissipation also. The amp was at height=4' with
nothing around sides or top. A small 12" fan (even with low airflow) on
the shelf would seem useful. Not blowing in, but blowing over/around, to
improve chassis dissipation and prevent hot spots on the intake air.
To say this more directly: I'm not sure that increasing blower speed is
the most effective way to spend dBA to control the amp internal temps.
(again, assuming the box has reasonable thermal transfer design inside)
2) What shack temperatures do people see? Do people expect their amps to
tolerate 100F intake air? (I'm expecting: yes?).
Note with 78F ambient, I'm measuring +69F exhaust air temp rise.
3) I see a lot of talk about tube temperatures, but the catastrophic
kind of RTTY related failures people post about don't seem related to
the tube? I read about coils/resistors/caps, or even plastic supports on
4) Does a 2:1 SWR increase any (I**2)R heat in the tank or other
circuits?. I'm wondering if the resistive load is not worst case.
5) My testing above was 1200W not 1500W.
6) I'm not worrying about operator error cases, like operating into the
wrong antenna etc.
7) If manufacturers built as many temp sensors into their products as
PCs do (drives, cpus, graphics, motherboard), we'd have a lot more info.
Just need to use these 1-wire sensors, going to a PIC then to a serial
port. (like the QK145 kit).
Come clean with a brand new shower. Click now!
More information about the Amps