TED-5000 2 MTU Weirdness.

[SunEagle]

Like I said on the TED-5000 forum.



Is it possible that all 3 got it wrong? I am not being sarcastic. I am truly asking. What more information do you guys need? Another picture with more detail? Let me know.

If it passed inspection there must fuses somewhere in the circuit.

Do what DanS26 says and carefully open the SQ D switch. Hopefully that is where the fuses are.

[DanS26]

I think the basic underlying issue here is that the line drawing is for a load side (ie distribution panel) connection, but somewhere down the line someone decided to do a supply side tap. So if some of the installation was load side based and some supply side based then things can get mixed up. Wrong wire sizes, missing fuses, etc.

Installers and inspectors can miss things when things change in the middle.

[RiggedCT]

If it passed inspection there must fuses somewhere in the circuit.

Do what DanS26 says and carefully open the SQ D switch. Hopefully that is where the fuses are.

Will do it Saturday night when I get home. Thanks guys. I added a post that had some color coded runs in the picture. something about a Mod having to approve it. Disregard it when it arrives. You guys have answered the questions.

[RiggedCT]

Will do it Saturday night when I get home. Thanks guys. I added a post that had some color coded runs in the picture. something about a Mod having to approve it. Disregard it when it arrives. You guys have answered the questions.

Seems There are fuses in the Square D disconnect.



Seems odd to me that the LOCUS data reporter only has one CT on one wire. Where as the TED has 1 CT for each. I have noticed a small difference in reporting from each device. It's only been one day.

23.315kWh TED-5000
23.021kWh LOCUS

[inetdog]

Seems There are fuses in the Square D disconnect.



Seems odd to me that the LOCUS data reporter only has one CT on one wire. Where as the TED has 1 CT for each. I have noticed a small difference in reporting from each device. It's only been one day.

23.315kWh TED-5000
23.021kWh LOCUS

When you have a 240V circuit with two hot wires and no neutral, the current in the two hot wires has to be equal.
When you have a three wire circuit (with 120V loads as well as 240V loads) you need to know the current in any two of the three wires to be able to calculate the total power.

[RiggedCT]

When you have a 240V circuit with two hot wires and no neutral, the current in the two hot wires has to be equal.
When you have a three wire circuit (with 120V loads as well as 240V loads) you need to know the current in any two of the three wires to be able to calculate the total power.

Thank you.

[SunEagle]

Thank you.

Looks like the mystery has been cleared up. Hopefully you get your TED system working.

[RiggedCT]

Looks like the mystery has been cleared up. Hopefully you get your TED system working.

Yes, We can mark this thread as solved.

[DanS26]

[QUOTE=

Seems odd to me that the LOCUS data reporter only has one CT on one wire. Where as the TED has 1 CT for each. I have noticed a small difference in reporting from each device. It's only been one day.

23.315kWh TED-5000
23.021kWh LOCUS[/QUOTE]

TED has a software adjustment to align the readings. You probably will want to align with the utility grade meter readings. On the meter above the switch.

[garW8]

140901-1004

RiggedCT:

You have apparently mostly solved your problem. However, I have some comments that may provide some useful information.

The carrier frequency for the TED 1000 and 5000 systems is exactly 125 kHz, within the tolerance of the cryatal oscillator in the MTU. This is in contrast to most statements that say the carrier frequency is 132 kHz.

In some of the TED literature they specify the 132 kHz value. Why? They should know what is in their product and how it works.

I have determined the actual carrier frequency three ways:

(1) Using the crystal oscilator frequency of the MTU, 8 MHz, and a integer divisor of 64 provides a result of 125 kHz. A divisor of 65 produces 123.1 kHz, 63 produces 127 kHz, 62 produces 129 kHz, and 61 provides 131.1 kHz. The only logical choice is 64 because 2^6 = 64. Any divider that is not 2 to an interger exponent is difficuklt to make. Thus, not likely used. 2 to the exponent 6 is 64.

(2) Adjusting a tuned LC resonant circuit to obtain maximum output from the carrier signal. Then determining the resonant frequency of that circuit, Result 125 kHz.

(3) Using a beat frequency measurement between the TED carrier frequency, and a sine wave oscillator, Again 125 kHz.


My suggestion for reliable PLC communication is:

This suggestion is based on the assumption that you want consistent
reliable data rather than the convience of plugging the receiver into any outlet in your house.

Pick a breaker in the main panel that is lightly loaded. Preferably with only TED components as the load. Connect the input side of an X-10 20 A filter to the breaker and neutral. Being in the main panel and with no other. or changing loads, provides the best estimate of voltage at the power company meter.

On the output side of the X-10 filter connect the MTU and its receiver (1000 system RDU, or 5000 system Gateway). and nothing else. This can be at least up to 250 feet of #14 Romex cable. This will provide essentially an error rate of 0 for a one MTU system. Functionally there is no difference between the input and output of the X-10 filter. The internal circuitry of the filter is symetrical.

Don't use the MTU red wire if one exists. This would require a second filter and provide no useful result other than the sum of the two 120 V phases would be used for power calculations.

If more than one MTU is used on the circuit, then there will be moderately long time periods when data errors occur. The data packets from the MTUs when this occurs should only show up as lost data at the receiver. The cause is overlapping transmissions from the multiple MTUs. This effect is certainly true for the 1000 system. I believe I also looked for it on the 5000 system. The reason is that the independent crystal oscillators in the multiple MTUs are not exactly the same frquency. So a transmission of one MTU gradually drifts thru the time period of another MTU, and overlapping at times.

If you expect to have the TED software combine the data from more than one MTU, then the multiple MTU devices must be on the same isolated communication path. In other words black and white wires from all MTUs must be wired together. The current transformers are connected to whatever circuits you desire.

To test the functionality of a current transformer make a test circuit consisting of a 100 W bulb and a split cable to the bulb. The split cable provides a means for you to clip a current transformer around just one wire to the bulb. 100 W incandescent bulbs are usually within +/- 2 W of bulb rating at bulb rated voltage.

You can measure power to individual 120 V loads by clipping one current transformer around a single wire to the load. 240 V circuits require two current transformers, one for each phase.

The 1000 system provides time resolution quantized to 1 second. The 5000 system does not. The 5000 system is 2 seconds and sometimes longer. Because TED fudges the data in the 5000 system it appears like there is 1 second data, but there is not. If you look closely at the 5000 1 second data you will see that consecutive data paired values are identical.

In the 1000 system data values are sent from the MTU once every 1 second. In the 5000 system data packets are sent every 2 seconds, and some times this rate slows down. Within a data packet there are not two values of the same measuremnt type (power, voltage, etc.).

A data packet transmission in the 1000 system occurs once per second and has a duration of approximately 0.1 seconds. This is a continuous transmission and encompasses 12 voltage zero crossings per transmission. 6 positive slope, and 6 negative slope. Two 1000 sysyem MTUs, when not overlapping, consume 0.2 xeconds of the 1 second period. The 5000 system is somewhat over 0.2 seconds per packet, and thus you can see why data is sent no faster than once every 2 seconds.

It is total nonsense to say that TED transmits only at zero crossings. It would take an excessive amount of time to do anything close to just at zero crossings, but it does interfere with other systems that transmit during some fraction of a 60 Hz cycle. Also mulyiple MTUs can interfere with each other.

.

[inetdog]

I have determined the actual carrier frequency three ways:

(1) Using the crystal oscilator frequency of the MTU, 8 MHz, and a integer divisor of 64 provides a result of 125 kHz. A divisor of 65 produces 123.1 kHz, 63 produces 127 kHz, 62 produces 129 kHz, and 61 provides 131.1 kHz. The only logical choice is 64 because 2^6 = 64. Any divider that is not 2 to an interger exponent is difficuklt to make. Thus, not likely used. 2 to the exponent 6 is 64.

(2) Adjusting a tuned LC resonant circuit to obtain maximum output from the carrier signal. Then determining the resonant frequency of that circuit, Result 125 kHz.

(3) Using a beat frequency measurement between the TED carrier frequency, and a sine wave oscillator, Again 125 kHz.

These results are just unacceptable. We all know that using three different methods must provide at least two incompatible results!

[bcroe]

140901-1004 RiggedCT:

You have apparently mostly solved your problem. However, I have some comments that may provide some useful information.

The carrier frequency for the TED 1000 and 5000 systems is exactly 125 kHz, within the tolerance of the cryatal oscillator in the MTU. This is in contrast to most statements that say the carrier frequency is 132 kHz.

In some of the TED literature they specify the 132 kHz value. Why? They should know what is in their product and how it works. I have determined the actual carrier frequency three ways:

(1) Using the crystal oscilator frequency of the MTU, 8 MHz, and a integer divisor of 64 provides a result of 125 kHz. A divisor of 65 produces 123.1 kHz, 63 produces 127 kHz, 62 produces 129 kHz, and 61 provides 131.1 kHz. The only logical choice is 64 because 2^6 = 64. Any divider that is not 2 to an interger exponent is difficuklt to make. Thus, not likely used. 2 to the exponent 6 is 64.

(2) Adjusting a tuned LC resonant circuit to obtain maximum output from the carrier signal. Then determining the resonant frequency of that circuit, Result 125 kHz.

(3) Using a beat frequency measurement between the TED carrier frequency, and a sine wave oscillator, Again 125 kHz.


My suggestion for reliable PLC communication is:

This suggestion is based on the assumption that you want consistent
reliable data rather than the convience of plugging the receiver into any outlet in your house.

Pick a breaker in the main panel that is lightly loaded. Preferably with only TED components as the load. Connect the input side of an X-10 20 A filter to the breaker and neutral. Being in the main panel and with no other. or changing loads, provides the best estimate of voltage at the power company meter.

On the output side of the X-10 filter connect the MTU and its receiver (1000 system RDU, or 5000 system Gateway). and nothing else. This can be at least up to 250 feet of #14 Romex cable. This will provide essentially an error rate of 0 for a one MTU system. Functionally there is no difference between the input and output of the X-10 filter. The internal circuitry of the filter is symetrical.

Don't use the MTU red wire if one exists. This would require a second filter and provide no useful result other than the sum of the two 120 V phases would be used for power calculations.

If more than one MTU is used on the circuit, then there will be moderately long time periods when data errors occur. The data packets from the MTUs when this occurs should only show up as lost data at the receiver. The cause is overlapping transmissions from the multiple MTUs. This effect is certainly true for the 1000 system. I believe I also looked for it on the 5000 system. The reason is that the independent crystal oscillators in the multiple MTUs are not exactly the same frquency. So a transmission of one MTU gradually drifts thru the time period of another MTU, and overlapping at times.

If you expect to have the TED software combine the data from more than one MTU, then the multiple MTU devices must be on the same isolated communication path. In other words black and white wires from all MTUs must be wired together. The current transformers are connected to whatever circuits you desire.

To test the functionality of a current transformer make a test circuit consisting of a 100 W bulb and a split cable to the bulb. The split cable provides a means for you to clip a current transformer around just one wire to the bulb. 100 W incandescent bulbs are usually within +/- 2 W of bulb rating at bulb rated voltage.

You can measure power to individual 120 V loads by clipping one current transformer around a single wire to the load. 240 V circuits require two current transformers, one for each phase.

The 1000 system provides time resolution quantized to 1 second. The 5000 system does not. The 5000 system is 2 seconds and sometimes longer. Because TED fudges the data in the 5000 system it appears like there is 1 second data, but there is not. If you look closely at the 5000 1 second data you will see that consecutive data paired values are identical.

In the 1000 system data values are sent from the MTU once every 1 second. In the 5000 system data packets are sent every 2 seconds, and some times this rate slows down. Within a data packet there are not two values of the same measuremnt type (power, voltage, etc.).

A data packet transmission in the 1000 system occurs once per second and has a duration of approximately 0.1 seconds. This is a continuous transmission and encompasses 12 voltage zero crossings per transmission. 6 positive slope, and 6 negative slope. Two 1000 sysyem MTUs, when not overlapping, consume 0.2 xeconds of the 1 second period. The 5000 system is somewhat over 0.2 seconds per packet, and thus you can see why data is sent no faster than once every 2 seconds.

That is a pretty extensive investigation. Been easier with an audio spectrum analyzer.
Guess they can run 125 KHZ if they want to; probably nobody bothered to update the
literature (cause they thought nobody would actually check). They could have used a
different crystal.

That X10 filter is a bandpass, I presume? A 125 khz resonant circuit with decent
Q is a bit of a challenge.

The next stage might be to detect the actual packets and decode them. Reminds
me of catching the commands of an infared remote control on my storage scope.
Every unit uses a completely different encoding scheme.

I didn't trust incandescent bulbs for accuracy checks. Instead, used a precision
shunt & power resistor load with true RMS meter to get voltage & current. Those
CTs were better than I anticipated. Bruce Roe

[garW8]

140901-2334 EDT

bcroe:

The X-10 filter is a multi-stage LC filter. A moderately high impedance shunting line to neutral, and a high impedance line-to-line. One does not want to shunt the PLC signal to neutral, and the desire is to keep the input to output impedance high at the carrier frequency to prevent carrier interference from transferring thru the filter. With this filter different TED systems can be on opposite sides of the filter and not interfere with each other. To build your own filter or have a knowledge of how the system works it is necessary to know the carrier frequency.

A precision high power resistor is the best way to check calibration as you pointed out. But for persons that don't have available such a resistor the incandescent bulb can be a useful quick check. Useful when looking for setup problems.

A TED system can provide relatively good accuracy on a unity power factor load or source. but has substantial problems as power factor drops. Try a 100 ufd load alone with a good polypropylene capacitor.

With 70 ufd at 120 V as the only load:
A Kill-A-Watt read 3.08 A, 1.6 W, 361 VA, and a PF of 0.01.
A TED 1000 read 28 W --- VA and PF are not available.
A TED 5000 read 30 W. 360 VA, and 0.83 PF. Current is not available but by calculation is the same as Kill-A-Watt.
The TED systems have enough error to prove that a power factor capacitor could reduce energy use, and a PFC capacitor will save no energy as read by the power company meter when connected close to the meter.

I have decoded the PLC data stream for both the 1000 and 5000 systems. Asynchronous serial communication at 1200 baud is used. There is one field for which I do not know the purpose. All power readings at the PLC point include sign.

,

[bcroe]

bcroe:
The X-10 filter is a multi-stage LC filter. A moderately high impedance shunting line to neutral, and a high impedance line-to-line. One does not want to shunt the PLC signal to neutral, and the desire is to keep the input to output impedance high at the carrier frequency to prevent carrier interference from transferring thru the filter. With this filter different TED systems can be on opposite sides of the filter and not interfere with each other. To build your own filter or have a knowledge of how the system works it is necessary to know the carrier frequency.

A precision high power resistor is the best way to check calibration as you pointed out. But for persons that don't have available such a resistor the incandescent bulb can be a useful quick check. Useful when looking for setup problems.

A TED system can provide relatively good accuracy on a unity power factor load or source. but has substantial problems as power factor drops. Try a 100 ufd load alone with a good polypropylene capacitor.

With 70 ufd at 120 V as the only load:
A Kill-A-Watt read 3.08 A, 1.6 W, 361 VA, and a PF of 0.01.
A TED 1000 read 28 W --- VA and PF are not available.
A TED 5000 read 30 W. 360 VA, and 0.83 PF. Current is not available but by calculation is the same as Kill-A-Watt.
The TED systems have enough error to prove that a power factor capacitor could reduce energy use, and a PFC capacitor will save no energy as read by the power company meter when connected close to the meter.

I have decoded the PLC data stream for both the 1000 and 5000 systems. Asynchronous serial communication at 1200 baud is used. There is one field for which I do not know the purpose. All power readings at the PLC point include sign.

OK, I still want to know if that is a band pass or band block (notch) filter, so I can understand
its function. Will this filter carry any 60 hz load power?

A while back I tried using an oil cap with TED, and I thought it did a pretty decent job of noting
a near zero power factor. Are you using oil, or a not so efficient chemical cap, or do I need to
go do that again? Bruce Roe

[garW8]

1140902-0756 EDT

bcroe:

The X-10 filter is a notch (band-stop) filter.

The X-10 PRO filter is Model XPF 20 A Wired-In Filter 120/220 V, from label.

Maximum input impedance (line-to-neutral) is at 155 kHz.

Input to output attenuation, voltage ratio, at:
120 kHz is 0.11,
130 kHz is 0.05, this is the frequency of maximum attenuation,
140 kHz is 0.09 .

Polypropylene is a very low loss dielectric and thus makes a very good capacitor at low frequencies. Dielectric constant is reasonable, breakdown voltage is good, and thin layers can make moderate size capacitors in moderate package sizes.

Above said filter is good for 20 A at 60 Hz and has low voltage drop.

.