MSi001 dongles. Frequency vs supply voltage.
(Jan 24 2014)
This page is a sub-page to The Mirics MSi001 with Linrad. The Mirics MSi001 chip is used in the FUNcube Dongle Pro+ and in the Logitec LDT-1S310U/J USB dongle. These two dongles are compared below.

Frequency stability vs supply voltage.

The dongles were tested in two different computers. The first dongle I had access to, the Logitec dongle, was nicely stable in my D5400XS computer. The FUNcube dongle did not work in the D5400XS computer. (ALSA could not find the dongle.) I moved to an older Pentium 4 where both dongles work. (Problems with USB drivers are not uncommon...)

As it turned out the frequency stability was very poor with random fluctuations of about 5 Hz for the Logitec tuner. Something had to be wrong and I suspected a contact problem. As it turned out, cleaning the USB connectors with isopropanol (like we do with SMA connectors) fixed the problem. Unfortunately I made no screen dump to show the error condition, but the fact that both dongles are sensitive to the supply voltage is demonstrated in figure 1.

Fig. 1. FUNcube PRO+ (left)and Logitec (right) dongles when an MP3 player is inserted first in a USB connector near the FUNcube dongle, then near the Logitec dongle.

The Pentium 4 motherboard has two groups of USB connectors. Four on the back panel and two more from a socket on the motherboard. With one dongle in each group a load was applied to one or the other group by connecting a MP3 player that would load by charging its battery. The modest load from the MP3 player shifts the frequency of the Logitec tuner by about 3.3 Hz if it is inserted in the same group and by 0.5 Hz if it is inserted in the other USB group. The frequency of the FUNcube PRO+ is affected by about 0.4 Hz in the same group and about 0.2 in the other group. When these dongles are used for modes like JT65 inadequate stability of the 5V line of the USB system could be the cause of degraded performance. To not degrade the performance, the USB voltage should be stable within 3 mV for the FUNcube and within 0.3 mV for the Logitec.

It might be a good idea to supply the dongles with 5 V from a separate very stable supply.

By inserting a PCI card that supplies four USB2 ports it was possible to run the FUNcube dongle in the D5400XS computer. The drive routine for the USB2 PCI card are evidently compatible with the FUNcube implementation of a USB soundcard contrary to the other USB systems on the D5400XS computer. Figures 2 to 4 show the frequency stability with the Logitec dongle connected to different USB ports on the computer. It is obvious that the ports differ.

Figure 2 shows that the Logitec oscillator has lower close range phase noise than the FUNcube. This is probably of no practical interest because all modes used at VHF/UHF allow for significantly more phase noise/unstability to accomodate for multipath propagation.

Figure 3 shows that there is an unstability in the USB supply voltage when both dongles are connected to the same PCI card. The FUNcube is less affected and the response seems to be delayed but it is to small for definite conclusions.

Figure 4 shows that whatever mechanism that causes the USB voltage to change by a couple of millivolts every 20 minutes is not always present. It also shows again that the close in phase noise is better for the Logitac dongle.

Fig. 2. FUNcube PRO+ (left)and Logitec (right) dongles on the D5400XS computer with the FUNcube on a PCI USB2 card and the Logitec tuner on a PCIe USB3 card.

Fig. 3. FUNcube PRO+ (left)and Logitec (right) dongles on the D5400XS computer with the FUNcube and the Logitec on the same PCI USB2 card.

Fig. 4. FUNcube PRO+ (left)and Logitec (right) dongles on the D5400XS computer with the FUNcube on a PCI USB2 card and the Logitec on a USB connector on the motherboard.